search this blog

Thursday, February 19, 2015

The Near East ain't what it used to be


Up for public comment at bioRxiv this week is this paper on the population history of the Near East, with a special focus on Armenians. Here's the abstract:

The Armenians are a culturally isolated population who historically inhabited a region in the Near East bounded by the Mediterranean and Black seas and the Caucasus, but remain underrepresented in genetic studies and have a complex history including a major geographic displacement during World War One. Here, we analyse genome-wide variation in 173 Armenians and compare them to 78 other worldwide populations. We find that Armenians form a distinctive cluster linking the Near East, Europe, and the Caucasus. We show that Armenian diversity can be explained by several mixtures of Eurasian populations that occurred between ~3,000 and ~2,000 BCE, a period characterized by major population migrations after the domestication of the horse, appearance of chariots, and the rise of advanced civilizations in the Near East. However, genetic signals of population mixture cease after ~1,200 BCE when Bronze Age civilizations in the Eastern Mediterranean world suddenly and violently collapsed. Armenians have since remained isolated and genetic structure within the population developed ~500 years ago when Armenia was divided between the Ottomans and the Safavid Empire in Iran. Finally, we show that Armenians have higher genetic affinity to Neolithic Europeans than other present-day Near Easterners, and that 29% of the Armenian ancestry may originate from an ancestral population best represented by Neolithic Europeans.

Unfortunately, the authors failed to even mention the main cause of what they're seeing; the massive influx of Ancient North Eurasian (ANE) admixture into the Near East. They included ancient genomes Oetzi the Iceman and La Brana-1 in their analysis, but not MA-1 or Mal'ta boy, the main ANE proxy.

MA-1 is a low coverage genome, and not easy to work with, but until better ANE reference genomes are sequenced, it simply can't be ignored in studies on the population history of West Eurasia. Here's why:


Above is my Fateful Triangle PCA. Note the eastern shift of the Islamic Near Eastern groups relative to their non-Islamic neighbors. Here are the relevant ANE ancestry proportions:

Anatolian Turks ~16.54%
Armenians ~15.48%

Iranians ~19.61%
Iranian Jews ~14.01%

Lebanese Muslims ~9.82%
Lebanese Christians ~7.14%

The differences aren't very dramatic, but they're consistent and, as per the PCA, hard to overlook. Indeed, the contrast would be even more obvious if we were to add to the list other exotic admixtures, such as East Asian, South Asian and/or Sub-Saharan.

If you're wondering why it is that Muslims generally carry more ANE than their non-Muslim neighbors, it's probably because the Islamic expansion had a homogenizing effect on the Near East, and it didn't have as much of an impact on the religious minorities in the region.

How and when ANE arrived in the Near East is still a mystery which can only be solved with ancient DNA. However, my bet is that most of it came after the Neolithic from the Eurasian steppe, the northeast Caucasus and the Altai, with the Indo-Europeans, Kura-Araxes people and Turks, respectively.

Citation...

Marc Haber et al., Genetic evidence for an origin of the Armenians from Bronze Age mixing of multiple populations, bioRxiv, Posted February 18, 2015. doi: http://dx.doi.org/10.1101/015396

See also...

First look at an ancient genome from Neolithic Anatolia

412 comments:

1 – 200 of 412   Newer›   Newest»
postneo said...

@david
"If you're wondering why it is that Muslims generally carry more ANE than their non-Muslim neighbors, it's probably because the Islamic expansion had a homogenizing effect on the Near East, and it didn't have as much of an impact on the religious minorities in the region."

Why so coy ? you could simply ask why do iranians to the south east of Armenians have more ANE than Armenians and turks ?

Its because they are farther away from the first farmers of the levant.

Let me ask further why do the turks not have more ANE than Iranians? Didn't they come from the Steppe?

Davidski said...

"Let me ask further why do the turks not have more ANE than Iranians? Didn't they come from the Steppe?"

Obviously, yes, they all moved to Anatolia from the steppe.

Now I have a question for you; what drugs are you injecting today?

Nirjhar007 said...

David is again Confused and trapped in his wishful dogma of Steppe Hypothesis....

postneo said...

@David
This is how the steppe hypotheses can work

ANE rich steppe populations ANE get to Iran and India via east caspian and not from Armenia.

What time frame you would you that in ?

Mike Thomas said...

I don't think you can try to predict which thoroughfare gene flow went thru mostly based on smidgins of differences in ane; esp given the differential impact of later processes and chance.

The east caspian c. 3000-2000 was likely a desert landscape .

postneo said...

@Mike
"mostly based on smidgins of differences in ane"

Sure but differences are not small. As you go south east of iran ANE is higher so its a steady cline from south asia to Turkey.

This needs some explanation. Either ANE needs to be better resolved with a southern or northern component or David has to postulate a super large steppe migration that brought ANE to India and Iran that dwarfs the Yamnaya migration.

Mike Thomas said...

Fair enough
So you proposing that the Iran region has ancestral ANE ?

DarthVadent said...

Hmm, it seems as if Armenians serve as good modern surrogates for Neolithic European genetic structure. But wouldn't this be due to the fact that unlike many of the ethnic group in the Middle East, that they lack any recent Sub-Saharan African and East Eurasian admixture?

Mike Thomas said...

Maybe
But I doubt they're wholly "pure"
And I don't think they're that close to oetzi and neolithic europeans , which suggests that there is a significant "Balkan Mesolithic" component in european farmers

Davidski said...

Armenians generally have too much ANE to be useful surrogates for Neolithic migrants to Europe. Samaritans have less ANE, so they're probably better, although not ideal.

But keep in mind that these new Haber et al. Armenians are different from those I have. They might have lower levels of ANE, at least some of them anyway.

Davidski said...

postneo,

ANE was obviously all over Central Asia like a rash before the Indo-European expansions.

Indeed, the Indo-Iranians started off north of the Caspian and moved around it in a clockwise direction to get to the Near East, so they probably picked up a lot of ANE in Central Asia, and thus possibly had even higher levels of it than the Yamnaya, even those from Samara.

It's hard to say when ANE first arrived in India, but it might have been there well before the Indo-Iranians. Nevertheless, we'll be able to track their movements into India despite this by using ancient and modern R1a-Z93 sequences.

Grey said...

"ANE was obviously all over Central Asia like a rash before the Indo-European expansions."

If that gap through the middle of the ANE map is a corridor cut by a later expansion then the range was pretty huge at one time.

https://drive.google.com/file/d/0B9o3EYTdM8lQOS1YODhvOWpWNzg/view

.

Wild speculating for fun but i'm imagining a mega Bantu expansion from somewhere around the Tien Shan to India, Iran and the Near East (but blocked from China) before the IE with the very wide fst between ENF and WHG

http://www.unz.com/gnxp/the-holocene-alien-years/

stemming from the ENF coming from far away without mixing with the HGs in between.

Alberto said...

Yes, Near Eastern were not what they used to be. This is the reason why I was suggesting the other day to use EEF as the best proxy for Ancient Near Eastern instead of any modern population from the Near East (including Bedouins). I do think it could make a difference.

Looking at the numbers that Chad was giving for Yamnaya: 30/45/25 (ENF/WHG/ANE), it implies that the Armenian-like population was 30% WHG, which seems quite off.

A 50/50 mix of EHG and Armenian-like should be more like:

Yamnaya:
30/32/38

Which is a 50/50 mix of:

EHG:
0/60/40

Armenian-like:
60/4/36

Which would make the Armenian-like pop a kind of Afghan Pashtun, sans the South Asian part.

Herders tend to move massively from areas that are becoming desertified to areas that are not, and not the other way around.

postneo said...

@Mike
"So you proposing that the Iran region has ancestral ANE ?"

David's proposing a double expansion a clockwise around the caspian and a separate yamnaya expansion from the same region.

I was actually thinking of ANE as a diffuse paleolithic component.

But the question in both these scenarios is why such a late presence in Europe ?

So a pre IE Iranian refuge for ANE seems a good bet. It avoids a double expansion and explains later late ANE in Europe.

Davidski said...

Postneo,

I never said there was a double expansion of anything.

What I said was that there were high levels of native ANE in Central Asia and most of Siberia, and the Indo-Europeans picked up a lot of it as they moved from the western steppe around the Caspian Sea and into South Asia and the Middle East.

My view is that the Middle East had 0% ANE until some started trickling in during the Neolithic from Siberia, before the big wave during the Bronze Age.

By the way, ANE levels have probably since plummeted in Central Asia and Siberia because of Turko-Mongol expansions, which carried a lot of ENA.

Alberto,

Just forget the Central Asian PIE hypothesis. It's pointless.

Mike Thomas said...

@ PN

"I was actually thinking of ANE as a diffuse paleolithic component."

That's what I think ; all the way to east Central europe.

@ Dave
..no ANE in ME..

Quite possibly , but the ME is a big, diverse place. I'd bet some ane in westbasian higlands from before the Bronze Age

Alberto said...

@David

"My view is that the Middle East had 0% ANE..."

Agreed.

"...until some started trickling in during the Neolithic from Siberia,..."

You think a population from Siberia, 5000+ YBP could have any impact in the Near East?

"before the big wave during the Bronze Age."

Big wave from Siberia? Isn't that an oxymoron? The Ket population today is about 1600 individuals.

"Just forget the Central Asian PIE hypothesis. It's pointless."

I'll mark those words.

Davidski said...

Yeah, I think the ancestor of V88 entered the Near East from Siberia with ceramic pottery.

The big wave, or rather big waves, of ANE during the Bronze and Iron Ages came from the Russian, Ural and Kazakh steppes, the north Caucasus and Central Asia.

And there's so much wrong with a Central Asian PIE homeland model that there's no point discussing it. I think that in the future, apart from ancient DNA, the main thing that will clinch the steppe hypothesis will be a more detailed study of the relationship between Proto-Indo-European and Proto-Uralic.

spagetiMeatball said...

Wasn't there a study a few years back that showed that christian lebanese were more closely related to europeans, west asians, and central asians, while muslim lebanese were more closely related to arabians and east africans. Relative to each other of course.

This post kind of contradicts that in a big way.

Mike Thomas said...

"Yeah, I think the ancestor of V88 entered the Near East from Siberia with ceramic pottery."

Ha ha . Ok . So wheres the "Siberian ceramic pottery" in the Middle East and africa ?

Mike Thomas said...

SpagBall
I think tho this study confirmed that Armenians. , Christian Lebanese , Sephardic Jews and Cypriots are nevertheless closer to europeans than Palestinians , Syrians and Muslim Lebanese

Mike Thomas said...

Dave you're on a roll tonight
"he main thing that will clinch the steppe hypothesis will be a more detailed study of the relationship between Proto-Indo-European and Proto-Uralic."


Where are the urals , and where is central asia ?

Davidski said...

sM,

That study from a couple of years ago said that Lebanese Christians were more similar to Europeans compared to Lebanese Muslims, and I think the Europeans the authors were mainly talking about were the HGDP Sardinians, who have very little, if any, ANE.

Lebanese Muslims have more ANE, so they're less similar to Mediterranean Europeans, especially Sardinians. But they also have more Sub-Saharan admixture than Lebanese Christians, which makes the more similar to Arabians in some tests.

Putting it all together, we can say that religious minorities in the Near East are more similar to Neolithic European farmers than their Muslim neighbors are, because the latter have inflated levels of post-Neolithic ANE and Sub-Saharan ancestry.

Mike,

It's there don't worry...

http://1.bp.blogspot.com/-EIhMueIHMqo/VOTwwhOb-tI/AAAAAAAAFBo/jpUWq-8U4cE/s1600/Capture.JPG

Nirjhar007 said...

@David
''the main thing that will clinch the steppe hypothesis will be a more detailed study of the relationship between Proto-Indo-European and Proto-Uralic.''
As far i see it will only fade it, we must remember That Zagros-Zarzian intrusions on Uralic area was happening since pre-History and it influenced Uralic vocabulary and also that Balto-Slavic has strong Uralic influence.
I think a more detailed work which quite nicely done by Sergeant showing PIE and Semitic correspondence, with more ancient Asian languages like Elamite etc is required and there is even phonetic similarity between PIE and some Munda languages in E India we are actually just scratching the surface...
Same for aDNA.

Mike Thomas said...

Dave ; thanks for that map link
Yes I agree that pottery spread from the Far East (Japan, Korea) to the west, splitting into africa and europe.

But what does that have to do with Siberian foragers, and R1b-V88, specifically ?

Matt said...

I think there maybe be a problem here with the idea of the Armenians specifically being very admixed and the dates on the basis of a large number of negative f3 stats, as there isn't really a similar comparison for similar stats for many populations, though I don't doubt Armenians are admixed. Also don't doubt the Near East has changed due to expansions from the south (whether or not those expanding guys and girls are actually properly called unadmixed ENF or ENF plus African).

Nevertheless, as with the Haak paper on Yamnaya and the steppe, we still get by far the strongest signal of admixture by f3 for Armenians when South Asian and Early European Neolithic farmers are combined, rather than anything BedouinB like or high in Yamnaya.

This is telling us something, what exactly is still hard to say - I think it is telling us that the strongest admixture in Armenians is actually a combination of European Early Farmers with an Ancestral North Indian which is somewhat different from a mere ENF+ANE / ENF+EHG signal* and also different from a mere ENF plus Yamnaya signal. (This might not necessarily tell us anything about the Indo-Europeans though, as its full possible they made little autosomal impact out of certain areas. Its possible steppe Indo-Europeans had very minimal autosomal impact outside of North Europe where their subsistence mode was reasonably competitive.) And EEF may actually be unadmixed Levantine early farmers - it is only a hypothesis that they are admixed between WHG and an earlier Middle Eastern farmer group, not explicitly proven.

Ultimately, time and adna will tell.

*ANI / West Asian probably will fit ENF + ANE best out of any combinations of the ENF + WHG + ANE clusters in the K8, but... I think that might be quite a "lossy" model if that were attempted in Haak residual modelling (just as BedouinB plus WHG plus EHG, let alone (BedouinB plus WHG plus MA1) is a much more lossy model in Haak residual modelling compared to the simple EN + WHG + Yamnaya model).

Chris Davies said...

Has anybody got any data on 'ANE' from a variety of African populations? Before anybody says it, I know that ANE is not 'African', however I would be interested to know. Also Yemenis and Omanis if anybody knows.

Davidski said...

Africans and Arabians don't have any ANE, unless they're mixed.

Roy King said...

@Davidski,
"http://1.bp.blogspot.com/-EIhMueIHMqo/VOTwwhOb-tI/AAAAAAAAFBo/jpUWq-8U4cE/s1600/Capture.JPG"
I hate to be snarky, but your map and the idea that pottery diffused from Siberia is just wingnut. The pottery of Japan in no way is akin to the pottery of say Sudan, let alone the Neolithic pottery of the Near East and Southern Europe. The map just denotes archaeological ages of pottery in the "Old World". Pottery production has many independent origins. The pottery of the indigenous poeple of the SW US resmbles Dhimini Ware from Greek Thessaly. Do you think there was diffusion from Thessaly to Arizona?

Chris Davies said...

"Africans and Arabians don't have any ANE, unless they're mixed"

So which African populations have ANE?

Grey said...

@Matt

"we still get by far the strongest signal of admixture by f3 for Armenians when South Asian and Early European Neolithic farmers are combined, rather than anything BedouinB like or high in Yamnaya."

Two sets of farmers imo: a mostly maritime Levant E set and a separate overland Taurus/Zagros G set who partially merged later. The E set having the Bedouin link and the G set the ANI link.

Grey said...

"So which African populations have ANE?"

Pastoralists near ancient gold fields would be my guess.

Davidski said...

Roy,

Both microblade technology and pottery moved into Eastern Europe and the Near East from Siberia. You can look this up I'm sure.

Various R lineages and some ANE might have come with.

Oaie Porc said...

on the PCAs the SE Euros (Northern Balkans - Yugoslavia, Romania, Bulgaria) come with no-abyss continuity to NW Europe (NW-SE continuum). can we make a single cluster NW europe + SE Europe from Norway to Romania? because they together stand apart from the other clusters - NE Europe, SW Europe, S Europe

Chris Davies said...

@ Grey -

My suspicion is that Omanis, Yemenis, Somalis, and Ethiopians will have ANE. And any African groups that have mixed with Cushitic speakers.

Matt said...

Grey: Two sets of farmers imo: a mostly maritime Levant E set and a separate overland Taurus/Zagros G set who partially merged later. The E set having the Bedouin link and the G set the ANI link.

Grey, seems possible that early agriculture could've developed at separate sites in Near East->South Central Eurasia, as there is a shared climate and crop stimulus, while peoples living there could have been genetically divergent (or maybe divergent and both affected by "Basal Eurasians", who knows?).

I don't know how settled that story is archaeologically. It doesn't sound crazy.

What do you mean by G set and E set? If it's y dna G is established in Neolithic Europe, so I would doubt it linking with an ANI group.

Roy King said...

@Davidski,
Yes, Mesolithic lithic technology moved from Central Asia to the Near East/Europe. In particular, trapezoidal lithics, found in Mesolithic contexts, even Franchthi cave in Greece, whose "fossil site" is Chateauneuf les Martigues near the mouth of the Rhone in France. (Castelnovian). Didier Binder, a French archaeologist, who specializes in lithics, once told me, with play on words, that the trapezoids are found from the Caspian to the Capsian (from Iranian Caspian to North African Capsian). I do believe pre-Neolithic lithics could trace early migrations of R1b to Europe and North Africa.
What you argued is continuity of ceramics which is patently untrue. The map shows Sudanese ceramics, which are pre-Neolithic, older than Near Eastern Neolithic ceramics which are after PPNB. The Sudanese/Sahelian ceramics are quite different from the Pottery Neolithic of the Near East and extremely different from the early ceramics of Japan. I think your steppe models are fascinating, but don't overplay your hand through over-generalization.

Davidski said...

Oaie Porc,

It's easy to split the NW European cluster from the SE European cluster using Component 1, which carries most of the variation.

That's because the main axis of genetic differentiation in Europe is north to south.

Chris,

All of the groups you just mentioned don't have any ANE, not even noise amounts.

That's why it's almost certain that most of the ANE entered the Near East after the Neolithic, after all of the major population movements from the Near East to NE Africa.

Grey said...

@Matt

"What do you mean by G set and E set? If it's y dna G is established in Neolithic Europe, so I would doubt it linking with an ANI group."

Just thinking out loud really.

If you look at it as a process it seems to me that a group that develops some level of farming/herding are likely to share the same y haplogroup (assuming patrilocality which may be a false assumption) and then expand dramatically afterwards.

So it makes me wonder if the y haplogroups that are commonly found among early farmers might initially have come from separate locations

e.g. one Levant, one Zagros, one somewhere else etc

even if they all eventually ended up in/on critical junctions e.g. Crete, Greece etc.

Grey said...

@Chris Davies

My guess is entirely based on if ANE is from the north then if anyone with ANE traveled all that way it would probably be for gold (or something equally valuable)

west

http://resize.over-blog.com/600x450-ct.gif?http://idata.over-blog.com/0/52/41/24/culture-et-artistes-sn-galais/empire-du-ghana.gif

east

http://en.wikipedia.org/wiki/Great_Zimbabwe

and if they left any dna behind it would be in people who lived nearby who used to feed the miners.

But all of those assumptions might be wrong.

Roy King said...

I think that the reason that the non-Muslim people of the Near and Middle East have less ANE than the Muslim people could be very simple. Most of the non-Muslim people are from highlands--Armenia/Lebanese Christians, even Assyrians who resisted conversion to Islam. Perhaps the ANE "sweep" hit the low lands more readily which later had a larger probability of being caught up in conversion to Islam being less remote. The Iranian Jews would have a Levantine component of course which also might decrease the ANE admixture.

Matt said...

Grey: So it makes me wonder if the y haplogroups that are commonly found among early farmers might initially have come from separate locations
It's possible - I doubt G links to an ANI wave though, as it seems very Early Neolithic in Europe, which should be your Levant wave.

Early agriculturalists in Iran / Zagros seems not unlikely though - http://www.livescience.com/37963-agriculture-arose-eastern-fertile-crescent.html / http://news.sciencemag.org/archaeology/2013/07/farming-was-so-nice-it-was-invented-least-twice. These people could be genetically divergent (less WHG like?) from the people living in the Levantine part of the Fertile Crescent, from the beginning.

Chris Davies said...

@ Davidski -

"Chris,

All of the groups you just mentioned don't have any ANE, not even noise amounts."

I'd be very interested to see actual ANE for Ethiopians, Eritreans, Somalis, Yemenis, and Omanis.

I believe that all Maasai have some ANE, anything up to 3.5% with an average of around 1.0%.
While all Saudis have some ANE, anything up to 12.5% with an average of around 6.5%.
Makes me rather doubt that all Horn of African populations, Yemenis, and Omanis have absolutely zero or almost near zero.

Grey said...

@Matt

"These people could be genetically divergent (less WHG like?) from the people living in the Levantine part of the Fertile Crescent, from the beginning."

That's one of the things I was wondering.

Chad Rohlfsen said...

Armenian isn't much better of a fit than abkhasians. Look at the f3 stats. 30-32% enf isn't inconceivable.

jackson_montgomery_devoni said...

@Grey,

Do not forget about Y-DNA haplogroup J. There is a very good chance that some J will turn it among early Neolithic remains from the Near East once they are tested along with E and G

jackson_montgomery_devoni said...

Turn up*

anthrospain said...

Obviously they are not what they used to be. Otherwise we wouldn't have Stuttgart with ~0% of ANE.

Simon_W said...

Note p. 46 f. in the Haak et al. paper:

"Modern Armenians have a
signal of admixture from the Yamnaya, as when we test f3-statistics of the form f3(Armenian;
Yamnaya, X) we find the lowest Z-score for f3(Armenian; Yamnaya, BedouinB) = -0.00296 (Z=-7.1).
However, the lowest Z-score of statistics of the form f3(Armenian; X, Y) involves the (X, Y) =
(LBK_EN, Sindhi) pair (value -0.00575, Z=-15.3), so the signal of admixture from the Yamnaya is
46
not the strongest one for Armenians. Moreover, as shown in SI 7, the Yamnaya have a negative f3-
statistic with (X, Y) = (Karelia_HG, Armenian). A negative statistic for both Armenians and Yamnaya
with each other as a reference population may suggest that a third (unsampled) population admixed
into both the Yamnaya and to Armenians."

To me this suggests that ANE didn't primarily come from the north to West Asia, but from the east, it passed through Iran. Presumably with R1b1, R1b1b and R1b1c.

PF said...

I have to really dig deeper into all this, but here are some initial thoughts for review before I go barking up the wrong tree.

Is it possible there was some sort of early "proto caucasian / west asian" population spread out between the eastern Med and northern Iran, associated with the development of very early agriculture and hg G2a? Perhaps also connected to Dienekes's West_Asian component in the k7b?

These very early G2a innovators diverged into several groups, some becoming EEFs (by subsuming a WHG-rich pop), some becoming coastal Levantines, Anatolians, some moving to Iran, and some morphing into "proto-Armenians", a particularly drifted (and admixed?) branch from the original famers?

This group was probably engaged in farming north of the Caucuses and were the first that the ANE-rich invaders absorbed. They took the women and killed the men as per standard MO, thus becoming partially "proto-Armenian" while maintaining their Y hg stability. The survivors fled back to the mountains and evolved into the high G2a Caucasian populations we see today.

Perhaps this is why Otzi is coming out as 5% "Yamnaya"?

Also that first split between "Anatolian" and "PNIE" in the linguistics paper from the previous post is telling. One distinction is the possible Uralic influence in the latter but not the former. Thus Anatolian could have been the language of the "proto-Armenian" or related group which survived the northern invasion, or at least was less influenced by it.

Yet, if the PIE was native to the general caucuses area, how to explain Tocharian? If this version of events holds then descendants of Tocharian-speakers should have some noticeable "proto-Armenian" ancestry. Do they?

Marnie said...

@Roy

"Most of the non-Muslim people are from highlands--Armenia/Lebanese Christians, even Assyrians who resisted conversion to Islam."

I agree.

It's easier to hide in the mountains. Also, people who live in alpine regions are quite self sufficient. Lots of water, open alpine areas for livestock, wood for building, mushrooms and tubers to augment meat diet. Also, caves to hide in.

Marnie said...

@PF

"Perhaps this is why Otzi is coming out as 5% "Yamnaya"?"

Primarily because of the Caucasus?

Probably not.


Shaikorth said...

"Yet, if the PIE was native to the general caucuses area, how to explain Tocharian? If this version of events holds then descendants of Tocharian-speakers should have some noticeable "proto-Armenian" ancestry. Do they? "

Possibly.

http://1.bp.blogspot.com/-LIug8kEPKW4/UFtC9XP-VsI/AAAAAAAAD1A/HTmsEscIGoI/s1600/MDLPwestasian.jpg

But again, needs to be verified with ancient DNA from Tocharians themselves.

Alberto said...

@Chad

"Armenian isn't much better of a fit than abkhasians. Look at the f3 stats. 30-32% enf isn't inconceivable."

Sorry, I didn't get your point. ENF is correct, the problem is with WHG.

If you have a 50/50 mix of 2 populations and one population (EHG) is 0/60/40 (ENF/WHG/ANE), with your numbers for the resulting:

Yamnaya: 30/45/25

It means that the other (Armenian-like) population is: 60/30/10. And 30% WHG for this Armenian-like population looks too high.

Seinundzeit said...

Matt,

I think the top score for Armenians (and Georgians) involving Sindhi+Neolithic European is a reflection of ANE+ENF admixture for them, since all South Central Asian populations are very rich in ANE admixture, while early Neolithic European farmers are at 0%. I can't see any other reason for why this is consistently their strongest signal of admixture, since they display no genetic connection to South Asia via countless other methods.

Basically, a combination of ANE+ENF is most parsimoniously modeled as a mixture between an ANE-rich South Asian population (that also has substantial ENF ancestry) and ENF-rich early Neolithic farmers, which is why Armenians and Georgians always get this result. It isn't due to actual South Asian admixture among Armenians and Georgians.

Alberto said...

@Seinundzeit

Yes, exactly. The original Armenians were probably something like Afghan Pashtuns but without the South Asian part (see my first comment above). The score would be better with this population if they had tested it, I think (more ANE, less SA than Sindhi).

When they (Armenians) moved to the Near East, they mixed with Near Easterners, so they got more NE and lost ANE.

Matt said...

Sein: Basically, a combination of ANE+ENF is most parsimoniously modeled as a mixture between an ANE-rich South Asian population (that also has substantial ENF ancestry) and ENF-rich early Neolithic farmers, which is why Armenians and Georgians always get this result. It isn't due to actual South Asian admixture among Armenians and Georgians.

Yeah, I'm not suggesting it is actually South Asian admixture as such. What it is is more like Europeans modelling as Karitiana plus Stuttgart before we had MA1 and MA1 plus Stuttgart before we had EHG and EHG plus Stuttgart before we had Yamnaya and MN, and finally like Yamnaya plus MN. And surely whatever the admixture is, it approximates to some degree like ENF plus ANE.

What is actually likely, I think, is that as Simon cites from the Haak paper a third (unsampled) population admixed into both the Yamnaya and to Armenians.

How closely this group actually approximates to ENF plus ANE is uncertain. How do you know how parsimoniously this additional ancestry fits as ENF (BedouinB minus African, essentially) plus ANE without anything like the residual modelling used in Haak? The fact that David has described the ANE component IRC collapsing when too many South Central Asians / Indians are included in the ADMIXTURE run suggests to me that this ANE plus ENF does give some fit to "West Eurasian" ancestry in South Central Asians, but not necessarily that good of one if ADMIXTURE forces it out when it has lots of South Central Asian related variation to explain.

I am not so willing to place all the eggs in the basket of ANE and ENF at the moment. The ENF component was based on just one model of Early Neolithic Europeans and BedouinB, based on estimates of African contribution to BedouinB, and assuming this is the only element distinguishing BB from unadmixed early farmers. I think more adna is needed.

Alberto said...

@Matt

"How closely this group actually approximates to ENF plus ANE is uncertain."

Yes, I agree. Until we don't get an ancient genome from this population we won't know exactly.

Before MA-1, this component was usually called "West Asian". Now we have ANE, and the West Asian looks like a mixture of ANE and ENF. But it's probably not that exactly. As you said, more adna is needed.

Seinundzeit said...

Matt,

For what it's worth, here are some f3 stats Everest did awhile back (these results are very old, as I recall these were done before David started experimenting with formal tests).

Georgians; MA1, Moroccan Berber -0.0024124 0.000928752 -2.59746

Georgians; Punjabi, Egyptian Copt -0.000942598 0.000548189 -1.71948

Georgians; MA1, Samaritian -0.000703919 0.000678017 -1.0382

Georgians; Assyrian, Tamil Brahmin -0.000452213 0.000676253 -0.668705

Georgians; Finn, Assyrian -0.000338311 0.000721905 -0.468637

Georgians; Finn, Egyptian Copt -0.00029834 0.000677243 -0.440521

Georgians; Assyrian, MA1 -0.000241591 0.000812226 -0.297443

Georgians; Baloch, Sardinian -0.000234238 0.000249709 -0.938043

Georgians; Assyrian, Punjabi -0.000231087 0.000737462 -0.313354

Georgians; Karitiana, Egyptian Copt -0.000227871 0.000791454 -0.287914

(I'm not posting the Armenian results, because a few samples that are clearly of partially European ancestry weren't filtered from the data-set, which skewed their results)

So your intuition is correct, we are seeing something broadly along the lines of ENF+ANE. In addition to the scores involving MA1, this is demonstrated via Punjabi+Egyptian Copt, Assyrian+Tamil Brahmin, Baloch+Sardinian, and Assyrian+Punjabi. In fact, all of their scores seem to involve an ANE signal (Finnish individual+Near Easterner, Karitiana+Near Easterner).

Interestingly, most West Asians have similar results.

Seinundzeit said...

Here are more results.

A northwestern Iranian, has some ancestry from the Caucasus:

NW Iranian; MA1, Moroccan Berber -0.00318997 0.00170137 -1.87495

NW Iranian; South Indian Christian, Bedouin_B -0.00305321 0.00107275 -2.84615

NW Iranian; Karitiana, Bedouin_B -0.00285608 0.00111719 -2.55647

NW Iranian; Karitiana, Egyptian Copt -0.00285599 0.001388 -2.05763

NW Iranian; Punjabi, Egyptian Copt -0.00274011 0.00114165 -2.40014

NW Iranian; Karitiana, Samaritians -0.00249887 0.00128246 -1.94849

NW Iranian; Assyrian, Papuan -0.00224157 0.00129189 -1.7351

NW Iranian; Low-caste North Indian, Bedouin_B -0.00223538 0.000901314 -2.48013

NW Iranian; HGDP Pathan, Bedouin_B -0.00218953 0.000854177 -2.56333

NW Iranian; Bihari Brahmin, Bedouin_B -0.0021516 0.000966372 -2.22647

Another individual from northwestern Iran, also has documented Caucasian ancestry:

NW Iranian; MA1, Moroccan Berber -0.00598099 0.00150709 -3.96857

NW Iranian; Pakistani Punjabi, Egyptian Copt -0.00440121 0.00126275 -3.48541

NW Iranian; Karitiana, Egyptian Copt -0.00424804 0.00146561 -2.89847

NW Iranian; Haryana Jatt, Egyptian Copt -0.00414197 0.00129155 -3.20698

NW Iranian; Punjabi, Egyptian Copt -0.00413679 0.00122132 -3.38714

NW Iranian; MA1, Samaritians -0.00400553 0.00145291 -2.7569

NW Iranian; Yoruba, English -0.00374901 0.00116761 -3.21085

NW Iranian; Egyptian Copt, Pulliyar -0.00369707 0.00129416 -2.85673

NW Iranian; MA1, Egyptian Copt -0.00367106 0.00160572 -2.28624

An Assyrian individual:

Assyrian; MA1, Moroccan Berber -0.00158426 0.0014992 -1.05674

Assyrian; Punjabi, Egyptian Copt -0.00104059 0.00120756 -0.861732

Assyrian; Finn, Egyptian Copt -0.000596125 0.00131188 -0.454404

Assyrian; Egyptian Copt, Anatolian -0.000505967 0.00121935 -0.414947

Assyrian; English, Egyptian Copt -0.000310552 0.00133702 -0.232271

Assyrian; Burusho, Sardinian -0.000273434 0.000863029 -0.31683

Assyrian; Sardinian, Punjabi -0.000221387 0.00104906 -0.211035

Assyrian;.Egyptian Copt, Anatolian -0.000171336 0.00128853 -0.13297

Assyrian; Sardinian, Sindhi -0.000111694 0.000867031 -0.128823

Assyrian; Anatolian, Iraqi -9.85667e-05 0.00127059 -0.0775755

For comparison, Pashtun results from the same run:

(HGDP, northwestern Pakistan-Afghanistan border)

Pathan; MA1, Moroccan Berber -0.00651518 0.00095069 -6.85311

Pathan; MA1, Yoruba -0.00452261 0.00131642 -3.43554

Pathan; MA1, Samaritians -0.00447944 0.000636407 -7.03863

Pathan; Karitiana, Egyptian copt -0.00417908 0.000783709 -5.33244

Pathan; MA1, NW Iranian -0.00405993 0.000887062 -4.57683

Pathan; Karitiana, Samaritians -0.00402346 0.000515634 -7.80293

Pathan; MA1, Kurd (Turkey) -0.00397863 0.000807771 -4.92544

Pathan; MA1, Palestinian -0.00392319 0.000412352 -9.51417

Pathan; Assyrian, MA1 -0.00390391 0.000797455 -4.89546

(Di Cristofaro Pashtuns, northern Afghanistan)

Pashtun; MA1, Moroccan Berber -0.00727688 0.000996164 -7.3049

Pashtun; MA1, Yoruba -0.00574758 0.00134948 -4.25911

Pashtun; MA1, Samaritians -0.00501548 0.000735112 -6.82275

Pashtun; Karitiana, Samaritians -0.00487108 0.000545825 -8.92425

Pashtun; Karitiana, Egyptian Copt -0.00475834 0.000839687 -5.66681

Pashtun; MA1, Anatolian -0.00470069 0.000915262 -5.13589

Pashtun; Karitiana, Sardinian -0.00469621 0.000438142 -10.7185

Pashtun; MA1, Kurd (Turkey) -0.00464017 0.000885136 -5.24232

Pashtun; Karitiana, Armenians -0.00461275 0.0003961 -11.6454

Pashtun; MA1, NW Iranian -0.00455773 0.000933166 -4.88416

(Myself)

Sein; MA1, Moroccan Berber -0.00730444 0.00141435 -5.16451

Sein; Karitiana, Egyptian Copt -0.00549403 0.00134709 -4.07845

Sein; Karitiana, Armenians -0.0053723 0.000983146 -5.4644

Sein; Karitiana, Samaritians -0.00534301 0.00111176 -4.80589

Sein; Karitiana, Bedouin_B -0.00521769 0.00107161 -4.86899

Sein; Karitiana, Druze -0.00508837 0.00097031 -5.24407

Sein; Karitiana, Moroccan Berber -0.00504788 0.00119 -4.24192

Sein; Karitiana, Sardinian -0.0049606 0.00100038 -4.95872

Sein; Karitiana, Georgians -0.00495467 0.000957117 -5.17666

Sein; MA1, Yoruba -0.00473739 0.00176348 -2.68639

Marnie said...

@Seinundzeit

"So your intuition is correct, we are seeing something broadly along the lines of ENF+ANE."

You don't need to spend hours and hours generating f3 stats and the like to see that "ANE", "farmer" and "WHG" autosomal components have a pattern extending from Europe to Kazahkstan. It's a pattern that is readily apparent from an admixture run on Eurasian populations, such as this one:

http://www.linearpopulationmodel.blogspot.com/2015/01/ancient-north-eurasians-north-south.html

f3 isn't buying you anything above other methods.

Regarding the ENF "component", is that Oetzi?

Based on the archaeological record of the Balkans, Oetzi's ancestors were likely almost entirely Mesolithic Europeans.

As to the "ANE" component, which Oetzi does have at about the 1 or 2 percent level, it's like due to many different processes back and forth, initially mostly between Asia Minor and the Balkans, and later through the Steppe.

With the exception of the high coverage ancient DNA areas such as Germany, it is impossible to determine the exact sources of the ANE component in South Asia, Asia Minor and Europe.

f3 is a complete waste of time without sufficient ancient DNA coverage.

you can go ahead and fool yourself, but most of the rest of us aren't fooled.

Seinundzeit said...

Marnie,

Lol, why so much scorn?

To be honest, I'm not going to disagree. The f3 stats I just posted are pretty much in line with what David has uncovered using ADMIXTURE in supervised mode ("West Eurasia K8"). Still, it's important to try several approaches, and see if the same general patterns are replicated across different methods.

Davidski said...

"Regarding the ENF "component", is that Oetzi?"

Marnie, first try and get your head around these concepts before criticizing.

Marnie said...

@Davidski

I've already asked you to define the "ENF" and "EHG" components, and you haven't done so.

Regardless, you just need to look at the clustering analysis in the Seguin-Orlando et al paper to see that Oetzi's ancestry is comprised of about 50% "HG", 50% "Middle East" and trace amounts of "Central Asia".

http://www.sciencemag.org/content/346/6213/1113.abstract

Based on archaeology of the Balkans in the last few years, it is extremely unlikely that 50% of Oetzi's ancestry is "Middle East" derived. It's much more likely that the "Middle East" component described in this paper is, in fact, associated with the Southern Balkan-Asia Minor Mesolithic.

I was talking to a coworker of mine this morning who is from Croatia. He gave me an earful about the nonsense that population geneticists have subjected people from the Balkans to in the last ten years.

Still at it, I see.

So, in the words of my coworker, "it won't last long" and "what a waste of money".

Oetzi's ancestors were mostly Mesolithic Europeans, with a bit of ancestry from Asia Minor. I don't need your crappy distorted stats to tell me that.

Regarding trying to extract a STEPPE ONLY HYPOTHESIS from f3 stats: with the current ancient DNA available, you're dreaming (and wasting your time.)

Davidski said...

I have done so. You just haven't bothered to read and understand what I've written. I'm not going to keep repeating myself. Use some initiative if you missed what I already said several hundred times on this blog.

Chad Rohlfsen said...

I'd say, go back to November of 2013 and try to catch up. Clear your calendar.

Marnie said...

@Davidski

Davidski, in Silicon Valley speak, if you can't define what you are doing in a few sentences, then maybe it's because you don't know clearly what you are doing.

So, what is "ENF", in a few sentences?

Davidski said...

Look in the top right corner of the first figure I posted in the blog post above.

Marnie said...

@Davidski

I just looked back through your posts in November on ENF.

Here's your definition, as of November, 2014.

""Near eastern" aka "ENF" is a proxy for early Neolithic near eastern farmers that Davidski created. Stuttgart is 68-72% ENF and 28-32% WHG. "

OK. So, like I said, your "ENF" looks a lot like the "Middle East" component mentioned in the Seguin-Orlando paper.

Based on recent archaeology, this component does not specifically come from the "Near East" but more likely is a derivative of the Balkan-Asia Minor Mesolithic.

Davidski said...

ENF peaks at almost 90% in isolated Middle Eastern groups like Samaritans, Yemen Jews and some Bedouins.

Balkan-Asia Minor Mesolithic? Haha.

Marnie said...

@Davidski

""Near eastern" aka "ENF" is a proxy for early Neolithic near eastern farmers that Davidski created. Stuttgart is 68-72% ENF and 28-32% WHG. "

The Bedouin are on the Seguin-Orlando run:

http://www.linearpopulationmodel.blogspot.com/2015/01/ancient-north-eurasians-north-south.html

Like I've said before, the light blue component on the Seguin-Orlando run is a Mediterranean component. And the Bedouin are a pastoral people, [who do play a form of the gaida shared with people from Asia Minor, by the way]. So, it's not very surprising that they would genetically look somewhat like other pastoralist people from the Balkans, Sardinia, Asia Minor, the Levant, Iran, Pakistan and Afghanistan.

Like I said, with regard to Oetzi, looks his "ENF" or "farmer" ancestors are derived mostly from the Balkan Mesolithic, given his proximity to the Balkans.

Chad Rohlfsen said...

Marnie,
Why would the Balkans be 40% Basal Eurasian before 7000bce, yet Hungarians had zero, quite a bit later. Not possible.

Davidski said...

The so called Mediterranean light blue component is not Mediterranean at all, but Middle Eastern farmer with a bit of WHG and no ANE. The reason it peaks in the Mediterranean today is because this region was less impacted by migrations from the steppe than the Middle East, and thus today generally has less ANE than the Middle East.

If you take out the excess WHG from the so called Mediterranean component and add a lot of ANE to it, you'll just get highland West Asian or Gedrosian.

Cpk said...

Anatolian Turks are 16% ANE whereas Southern Italians / Cypriots are 8%. Anatolian Turks are also ~10% Central Asian. If ANE came with them Central Asian Turks who invaded Anatolia were extremely high ANE.

Davidski said...

Here's what I said in the blog post above:

"How and when ANE arrived in the Near East is still a mystery which can only be solved with ancient DNA. However, my bet is that most of it came after the Neolithic from the Eurasian steppe, the northeast Caucasus and the Altai, with the Indo-Europeans, Kura-Araxes people and Turks, respectively."

Anatolians were impacted by all of these migrations IMO.

Cpk said...

I am saying pre Turkish Anatolians were probably ~8% ANE like Cypriots/S.Italians.

Marnie said...

@Chad,

"Why would the Balkans be 40% Basal Eurasian before 7000bce, "

I think it's already been established that the "Basal Eurasian" idea is bogus.

"yet Hungarians had zero, quite a bit later."

Hungary is not part of the Southern Balkans. Early Neolithic ceramics between Hungary and the Southern Balkans are different.

Maybe if you weren't wasting so much time doodling around with f stats, you could read a few recent archaeology papers on the Balkans.

Anyway, I think its funny how you guys seem to keep mixing up all these Balkan countries.

Davidski said...

Cpk,

Wouldn't pre-Turkish Anatolians be more similar to Armenians than to Cypriots and South Italians?

Cpk said...

I think they would be closer to the Greek-related populations. (except maybe Eastern Anatolians) Modern Greeks have Slavic admixture so Cypriots and South Italians would be better examples.

Chad Rohlfsen said...

Alberto,
You're taking the f3 too literally. Just wait and see.

Marnie,
It's not bogus. You clearly don't understand. I'm not mixing up the Balkans. If you remember, I had to correct you on which cultures were in certain modern modern countries. But, carry on with misinformation.

Davidski said...

These Turks are from central Anatolia, so their ancestors were probably closer to Armenians than to Cypriots and Greeks.

Western Turks are generally more similar to Greeks because they have less Turkic ancestry, but they probably have less ANE than Armenians.

Marnie said...

Chad,

I'm not going to be taking geography or anthropology lessons from you, that's for sure.

Cpk said...

If we can find some Anatolian Turks with no Central Asian admixture we can find out the pre Turkish Anatolian ANE levels.

Marnie said...

@Chad

"But, carry on with misinformation."

Based on various conversations I've had with people (Armenians, Greeks and Croatians, for instance) I'm not the only one who is fed up with the junk you've been spoon feeding the gullible for the last ten years.

Anyway, as my Croatian co-worker told me this morning, "It's only a matter of time."

And, yes, he's right, these bogus pop gen studies are "a waste of money."

Mike Thomas said...

But surely marnie has a point about the possible contribution of balkan mesoliths. EEF do not cluster with modern bedouins, Palestinians, etc but are clearly distinct and are rather closer to modern South europeans and non-Arab near easterners

Davidski said...

Marnie's point is pointless.

Balkan foragers did contribute gene flow to early European farmers, and this is why early European farmers appear to be partly Western European Hunter-Gatherer (WHG).

What this means though is that Balkan and Western European hunter-gatherers were very similar, and not Anatolian-like or even Greek-like, which is what Marnie would like to see.

Mike Thomas said...

Yeah Yr likely right ;
Though I do find it interesting that virtually all of Europe (balkans to NW europe) falls under a near homogeneous WHG; which SHG stand apart . Must be due to early ANE : EHG admixture

Krefter said...

@marnie,

"Based on various conversations I've had with people (Armenians, Greeks and Croatians, for instance) I'm not the only one who is fed up with the junk you've been spoon feeding the gullible for the last ten years."

What a handful of individuals say doesn't matter. The DNA speaks for itself.

Stop lying about people on these blogs. No one here is a conspirator.


Will you finally open your eyes and realize several of us get fed up with you because you're arrogant and believe in things as worth as ancient aliens(Chad being a lier).

Simply change those things.

@Marnie,
"Anyway, as my Croatian co-worker told me this morning, "It's only a matter of time."

And, yes, he's right, these bogus pop gen studies are "a waste of money.""

So, sampling the genome of 69 pre-historic humans is simply a "pop" study.

I don't understand why you are so fond of Balkan and near eastern people. Haak 2015 doesn't dislike them in any way.

Simply being critical of this paper and calling it a "pop" study, isn't an argument. You don't give evidence you just give doubt.

Marnie said...

@Krefter

I'm sure someone is paying you to generate this garbage.

Anyway, like I said, it's only a matter of time.

Good for you though, should be worth a few years of gravy.

. . . but, won't be good for your reputation long term, if you care at all.

Davidski said...

Marnie, Krefter's a 17-year-old kid, and he's currently got a much better grasp of the main concepts here than you do.

For your own sake, try and fix this sooner rather than later.

Marnie said...

@Davidski

"Marnie, Krefter's a 17-year-old kid, and he's currently got a much better grasp of the main concepts here than you do."

We don't know who Krefter is. He doesn't have an identity so he could be a flea, as far as I'm concerned.

Yes, Krefter seems to have a very good grasp of the *AGENDA* you guys have all agreed to churn out, courtesy of your funding masters.

"For your own sake, try and fix this sooner rather than later."

Hey baby, I don't have a pop gen reputation to protect, so I have nothing to "fix."

Grey said...

@jackson

"Do not forget about Y-DNA haplogroup J. There is a very good chance that some J will turn it among early Neolithic remains from the Near East once they are tested along with E and G"

Yes. I'm thinking if the different farmer haplogroups were initially separate then the E strand would be easiest to isolate but that's just a guess.

Chad Rohlfsen said...

It's a big conspiracy. We've already implanted chips into everyone's mind. Soon, we will rule the world... the aliens made us do it.

postneo said...

@Mike
Is ANE in south Asia and siberia shared from paleolithic times?

Maybe ... then why did it not spread to Europe? Why are Alaska and Bering straits more attractive than Europe?

Were there geological barriers? We have caspian and baikal seals found in really odd places. Basically these are stranded arctic species.

Perhaps retreating ice sheets formed a network of frozen lakes unattractive to humans for quite a while.

Mike Thomas said...

Postneo
Why do you think it hadn't spread to europe ?
it's in Mesolithic Karelia; small amounts in SHG; and likely also those parts of eastern europe not yet sampled- like the South baltic and East Carpathian

Chad Rohlfsen said...

David,
Check out S8.11 (b). That is the direction that I'm leaning.

Krefter said...

Maybe EHG is WHG+relative of MA-1. It would be strange if MA-1 was part something which as far as we know arose 15,000 years after he died.

Davidski said...

Yeah, ANE was in the east Carpathians. That's why none of the ancient Hungarian genomes have it; the ANE-rich folks obviously couldn't climb over the mountain.

Chad Rohlfsen said...

EHG fails as a mix of WHG and MA-1. They scatter all over the damn plot, with no correlation. It's not a guarantee, but points more towards the section of the paper I favor.

MA-1 might be the 3rd population of the "Near East" part and Yamnaya.

Mike Thomas said...

Dave
"Yeah, ANE was in the east Carpathians. That's why none of the ancient Hungarian genomes have it; the ANE-rich folks obviously couldn't climb over the mountain."

Either that; or they were scared of water - hence they stood still at the Dnieper river for 20000 years

Mike Thomas said...

Chad please elaborate / clarify ?

Chad Rohlfsen said...

Start at Section 8, of the Supplementary Info, through Section 9.

Mike Thomas said...

Ah so you're thinking that it is WHG that is the more admixed population cf EHG ?

postneo said...

@Mike
I was alluding to the fact that Ma-1 did not reach europe for another 12000 yrs

What stopped them. It must be the geographical barriers that opened up only during the mesolithic.

Mike Thomas said...

I think Malta was a failed dispersal North
What stopped them was their own lack of success :)

Traditional archaeology has placed the origins of russian plain upper Palaeolithic sites to Central Europe - Balkans . In this regard, then, it might be not surprising that Kostenki was Hg C.

Thus Hg R dervived groups mihjt have arrived later, from the southeast; but obviously I'm in the realm of speculation here

V.R. said...

@ Krefter,

You're only 17? Wow, that makes me feel dumb.

@ Marnie,
Put a cork in it.

Alberto said...

@Postneo

Regarding ANE from Paleolithic times I think it could just have to do with evolution/drift. We're talking about long periods of time and long distances (with different environmental conditions), and as populations move in space and time they change (even without mixing with some other population).

WHG could well share an origin with ANE, but as they moved west and stayed there for several 1000s of years (till the Mesolithic) they evolved into what we call WHG.

This could also explain why EHGs don't appear to be admixed. They could be in an intermediate point in evolution between ANE and WHG.

All speculation, of course, but just proposing a different possibility.

Seinundzeit said...

David,

Do you think it would be worthwhile to try computing fst distances between populations in your data-set+some of the aDNA samples? Similar to what they've done in Haak et al., if you feel it would be worth the effort.

Mike Thomas said...

Yeah Alberto
I agree that the all the populations with ane aren't bound by one unifying demographic history

Davidski said...

Sein, I don't have anything installed that does Fst calculations on SNP data. If you've seen something that is quick and easy let me know.

Grey said...

@postneo

"then why did it not spread to Europe? ... Basically these are stranded arctic species"

http://www.marietta.edu/~biol/biomes/alpine.htm

Roy King said...

@Davidski
"Marnie's point is pointless.
Balkan foragers did contribute gene flow to early European farmers, and this is why early European farmers appear to be partly Western European Hunter-Gatherer (WHG).
What this means though is that Balkan and Western European hunter-gatherers were very similar, and not Anatolian-like or even Greek-like, which is what Marnie would like to see."

I am curious, Davidski. What evidence leads you to conclude that the hunter-forager-fisher populations of Franchthi Cave in Greece, Crete and Western Sicily are genetically close to the WHG? After all EHG is different from WHG. Absence of evidence is not evidence of absence as archaeologists are fond of saying.

Mike Thomas said...

I'd concur at that possibility . Indeed SHG is also different . What are homogenous are Spain, NW Europe and the Carpathian basin- which is not surprising given that the former two naturally derive from the latter ; whist we'd expect greater diversity within the Balkans propper.

Davidski said...

Roy,

Most of the non-Basal Eurasian input in modern Near Easterners is very WHG-like, so much so that it's difficult to differentiate from WHG. So why would foragers in southern Italy and the southern Balkans carry a component that was distinct from what was all around them in Europe and the Near East? It wouldn't make any sense.

Btw, those sites you mention are obviously on the fringes of Mediterranean Europe, so it's possible that the remains there will show some Basal Eurasian admixture from the Near East or North Africa. But we already know from the ancient DNA from the Hungarian Plain that everyone there until the Copper Age was WHG-like or a mixture of WHG-related ancestry and EEF. So I don't expect any surprises from any mainland Balkan sites.

For whatever reason Marnie basically wants to make Oetzi the Iceman a native European forager, and that just won't work.

ryukendo kendow said...

Hi guys, I read the paper one last time, some more thoughts. Posting them here as the other thread is too long.

The f4 modelling is managed by Lazaridis in the paper. Sorry for misattribution, will refer to it as Laz modelling from noww on lol.

1)

About the whole admixed/non-admixed for EHG issue, saying EHG is intermediate between ANE and WHG is as informative as saying that Spanish are in between Mexicans and German (Which they are), or that Europeans are between Bedouin and Dai. It doesn't really explain much.

I know I'vve been harping on about how good the f4 stats are. Just as a recap, the laz modelling works by finding out the differential rel of 2 pops A and B to 15 outgroups, then seeing if C is in between them. If C is, then for some comparisons of A and B and C with a pair of outgroups, C will seem closer to A and further from B, while for other comparisons C will be closer to B and further from A, and the C-B and C-A distance should seesaw, growing and shrinking oppositely in a negatively correlated way; there should be no instance where C gets closer or further from both A and B at the same time. If there are such cases, either 1) A and B are too close together, or 2) C is not a product of A+B. It bears mentioning that laz modelling is insensitive to high drift, high homozygosity etc. etc.

Two other quotes from the paper.

"To gain understanding on why this method works, we examined the particular
statistics plotted in Fig. S9.12 and noticed that the pattern of linear anti-correlation was most strongly
affected by Native American and North Asian populations"

"The statistics involving Yamnaya
and LBK_EN have consistently opposing signs (thus Corded_Ware_LN is intermediate between these
populations), and involve comparisons with the Native American Karitiana and the Chukchi and
Eskimo populations from northeast Siberia. The inclusion of these populations (who have “Ancient
North Eurasian” ancestry2) can discriminate between LBK_EN and Yamnaya, as Yamnaya share most
drift with them, LBK_EN least, and Corded_Ware_LN intermediate."

ryukendo kendow said...

So they indeed do confirm that differential relationship to Siberian and NAm genomes strongly distinguish EHG and ANE ancestry from all other ancestry, including WHG ancestry. We also know that laz modelling can reveal Motala as EHG+WHG, a.k.a EHG and WHG are sufficiently differentiated w.r.t outgroups to revveal Motala as admixed. Putting these together, as EHG is intermediate between WHG and ANE, why would Laz modelling succeed in distinguishing EHG and WHG, but not ANE and WHG? Especially as differential rel. to Siberians and NAms mediates both? I think this kinda weakens the author's proposal that 'WHG and ANE are not sufficiently differentiated w.r.t the outgroups, and cuts 1) out, leaving 2).

The fact that outgroup changes alter the results so little (fig s19.6) also suggests that the result is robust.

The authors do say the relationship betwween EHG-WHG-ANE is complex, but they also say EHG emerges as unmixed in f4 stats, EHG is equi rel. to NAms as MA-1, tree-building with EHG as WHG+ANE fails. Of EHG-WHG-ANE, onee of these must be admixed; EHG admixed fails, so either MA-1 or WHG is admixed. Of the largest trees, all the best fits show WHG as admixed between a clade (ANE-EHG-NAms' portion of ANE) and an Outgroup.

I've whipped out a pic in ms paint that shows the one of the trees (there are three) that explains all the data; this is the one that best fits geography.

http://imgur.com/Czrp5mw

In this model EHG shares more alleles with WHG than ANE does, because EHG contrib to WHG. EHG shares more alleles with ANE than WHG does, because EHG split from ANE later than WHG did. EHG is equi with NAms as MA-1 is, because both ssplit from the ANE that went into NAms at the same time. And all the pops are arrayed east to west, without criss-crossing migrations are supposed to leave 0% trace behind, and without positing that Mal'ta's descendants were crushed under the wave of Karelians that reached America, leaving 0% Mal'ta-like contrib to Americas.

More things that make me think this: 1) EHG and ANE behave quite differently in Europeans in stats, peaking in different places. 2) Distortions. Using the f3 stats in ADMIXTUREGRAPH, 40% ANE+ 60% WHG admix proportions for EHG are inferred as the best estimate to minimise residuals, but for laz f4 stats modelling, least residuals are reached when Europeans are 25% or less EHG. if EHG is just ANE + WHG, this implies the most ANE pop in Europe is <10% ANE and all other europeans less, which is a massive levvel of distortion.

@ Sein

I agree with you that ANE as centered in Mal'ta is unadmixed w.r.t the genomes that we have now.

I disagree that ADMIXTURE can help us discern the direction of gene flow. E.g. When Mal'ta is included, E. Euro, Cauc, indus Valley and ASI pops are seen to contribute to Mal'ta when the opposite is true, and conclusions from formal stats are what gave us the info needed to supervise ADMIXTURE to produce the K8, not info from ADMIXTURE itself. I think as ADMIXTURE makes its clusters around groups of pops which differ similarly from the rest, it seems to attribute gene flow based on which pops you have more of, and those would seem to contribute to everyone else since they get their own component.

I think clines can produce fuzzy trees, but it doesn't really apply here, because trees where EHG are admixed are nevvertheless 'hard' in that Mal'ta and Mal'ta-likes must contribute 0% to NAms, and 100% of contribution to NAms must come from Karelian-likes, and the closer the figure is to 0% and 100% the better the fit even though it is still not that good at that extreme.

ryukendo kendow said...

2)
I made a prediction that West Asian ancestry will distinguish between IE contribution and EHG survivval in Europe, with EHG survvival in East Europe inflating the Yamnaya estimates there in whatever modelling techniques used. The inflation takes place because CW has i) more yamnaya ancestry than any present-day euro, and ii) more EHG ancestry than any present-day Euro. The main feature--the high EHG in CW--would cause them to come closer to E Euros whatevver the actual composition of the ancc of E Euros, as these Euros differ in the same way from other Euros (high EHG) as CW does from all present-day Euros (also high EHG), but this tells us little abt where the EHG comes from. It seems like there is some evvidence for this:

a) EHG has top drift with E Euros in drift f3 stats, while the IE pops incl. Corded Ware, Unetice, LNBA all havve top drift with NW Euros, Lithuanians excepted. This is in fitting with the higher amt of ENF ancestry from the NE/Cauc inferred for NW Euros, which would be the case if they had higher Yamnaya.

b) Laz modelling, which givves Estonian 52% yamnaya, also givves the adjacent finnish 55% Yamnaya, when the true figure is prob much less; the high yam figure is prob due to high EHG. Suppose the finnish were only 30-40% yamnaya or less, given their uniparental markers, and that the 10-20% residual is attributed to EHG, and that this effect is spread around NE Europe to their adjacent pops.

c) Here is a plot of the f3 shared drift stats between Corded Ware and EHG done using extremely basic image editing in 5 min:

http://imgur.com/kpNrdcX,k73e5qN#0
Blue are IE aDNA, Green HG aDNA, Black NW Euros, Brown NE Euros. Lithuanian are red.
CW shared drift shifts pts north, EHG shared drift shifts pts east. All the IE aDNA are shifted north, HG aDNA shifted east, NW Euros north, NE Euros east.

E Euros, e.g. lithaunian, are only slightly less EHG than IE aDNA, but much less CW than IE aDNA, while NW Euros show a more proportional drop in both.

The above pattern is not very strong, prob bc f3s are very 'fragile', but a more striking pattern is seen in f4(Test, CW, Yamnaya, Chimp) vs f4(Test, CW, Karelian, Chimp).

http://imgur.com/kpNrdcX,k73e5qN#1
EHG shifts south, and Yamnaya shifts east. NW Euros Orange, NE Euros Green, all other Euros black. It seems most euros incl NW euros have closeness with EHG mediated 100% through CW; aka NW and other Euros have as much EHG affinity as would be expected from their CW contribution. But East Euros have 'more' affinity with EHG than their affinity with CW would suggest, implying that there is 'more' EHG in their genomes on top of what CW gave them. The pattern is very in fitting with Geography--Ukrainians and czechs are very slightly 'inflated', Belarussian and Russian intermdiate, the most are Lithuanian and Estonian.

Putting this together with the Laz modelling which gives equal yamnaya to a swathe of N Euros, this means that, if some portion of 'Yamnaya' ancestry in East Euros actually comes from EHG survvival, perhaps... East Euros are less Yamnaya than NW and NC Euros.

ryukendo kendow said...

This is an extremely counterintuitive result, but I've always thought that the neol side of Yamnaya seems to come from the NE, and that this direct NE ENF influence is different from the NE influence from EEF tthrough the meditterranean, and that this NE influence reaches a trough in NE Euros, which suggests a dip in IE ancestry there. While this might not be geographically pleasing, consider that the CW andd Yamnaya come from more than 3kya ago, and 3kya ago africa was not bantu or 'black', central asia was not turkic or 'asian', SEAsia had no tais, S China had no Chinese, Madagascar no Malagasy, MENA no Arabs etc. So a second pop turnover there is not an impossible prospect, esp considering that it was so common in all areas for quite some time, and the agricultural frontier has moved deep southwards from the forest with the exp of the slavs.

In any case, a Laz model, or an ADMIXTURE analysis, containing WHG and both EHG and Yamnaya as sources of eastern acestry would make this clear quite soon.

Apologies for the poor-man's data analysis.

ryukendo kendow said...

3) Something weird is happening to Iceman.

We already know that Iceman scores a bit 'east' in PCAs; here Iceman even gets a bit of Yamnaya. The rank modelling that takes middle neol farmer genomes as a two-way contrib of WHG+early neol genomes does the worst for Iceman.


@ Davvidski

I don't think words like 'native', 'northern' are that helpful. This is what the oldest and thus the most 'native' northern man looked like:

http://i.kinja-img.com/gawker-media/image/upload/s--A05nTlB9--/c_fit,fl_progressive,q_80,w_636/tdjgnb58f3aknofmaqei.jpg

In fact, his limb proportions are akin to those in papuans and africans today and quite far from those of europeans and asians now, which are much shorter. And he is a basal west eurasian and got replaced shortly after by other intrusive pops from the West Eurasian clade coming from the same place where he originally came. So the repop/replacement from which most Euros get ancestry today probably wasn't from the north.

The clade (ANE-EHG-NAm's portion of ANE) seems to have gotten its start as an old population in SC Asia in any case, and, as Kostenki had none, it seems to arrive northwards from the south later than any part of WHG did. And also by the fact that 'ANE' as convventionally defined is found in levels equal to those in Lezgins even in South Indian Tribals with ~0% IE ancestry.

"We already know K was in Western Siberia 45,000 years ago, R in South Siberia 24,000 years ago, and both R1a and R1b in pre-Neolithic Eastern Europe. But that's all we really know."

This seems to suggest a picture of continuity in Siberia--and to use that to cast aspersions on present-day distr and to suggest expansion from Sibera--when in fact from aDNA we know there is really little case for any such exp. Ust Ishim is basal to even Papuans, his autosome was replaced, and he left no descendants; Kostenki carries something basal to all west Eurasians today, his autosome similarly replaced and left no descendants; Mal'ta similarly had his autosome replaced and left no descendants along his Y-line, although his cousins occupying the same phylogen position as he, and who even preserve something similar to his autosome, can be found quite far away today.

This seems to suggest a history of repeated repops of Siberia from the south, and the Y-chrom findings have to be considered in light of this kind of autosomal history. Also from modern autosomal we know places like India has seen pop continuity through ASI seen through split of Onge ~50 Kya, while places like Siberia has seen repeated replacement instead. The genomes from 45 to 36 to 24 to 9 kya are all drastically different.



@ The disc about Armenians
It seems that Agris, when they spread east and NEast from levant+Mesop, encountered an ANE-rich HG pop in SC Asia and the caspian area instead of a WHG-rich one, and that this pop had some influence in the ME and Cauc.

Davidski said...

rk,

I have to say that the Bedouin-Dai comparison is really not very useful.

It actually makes a lot of sense to model EHG as WHG/MA-1, or MA-1 as EHG/ANE. But it makes very little sense to model Europeans as Bedouin/Dai.

Shaikorth said...

Indeed, Sardinan-Dai and Basque-Dai would be better comparisons although still don't tell the whole story.

Mike Thomas said...

RK
That is similar to what I have been, except you've based on more actual analysis of the data; and I did it on sheer analysis of the patterns I see and what I expect from archaeology

As a minor point this ; see how little West asian corded ware has c.f yamnaya; which makes the supposition that it dervived 75% from yamnaya less likely.
Moreover ; we'd need to see if any west asian arrived also toward the steppe and Central Europe ; hence the west , via balkan-Anatolian route as well.

There is the possibility to consider if EHG arrived in Western Europe via CWC whilst R1b did via balkans

Mike Thomas said...

"This seems to suggest a history of repeated repops of Siberia from the south, and the Y-chrom findings have to be considered in light of this kind of autosomal history. Also from modern autosomal we know places like India has seen pop continuity through ASI seen through split of Onge ~50 Kya, while places like Siberia has seen repeated replacement instead. The genomes from 45 to 36 to 24 to 9 kya are all drastically different."

Ie R1 didn't develop in EE (?)

Mike Thomas said...

& RK

" this means that, if some portion of 'Yamnaya' ancestry in East Euros actually comes from EHG survvival"'

Something I also argued since the paper came out

Davidski said...

The Yamnaya and CWC horizons between them covered practically all of Eastern Europe. Good luck trying to split the Yamnaya/CWC EHG from the other EHG, whatever that actually means.

ryukendo kendow said...

@ Shaikorth @ Davidski

The only reason why it makes sense more in the latter case is because we havve circumstantial information that tells us Bedouin + Dai = European is not a realistic scenario. But in the language of pure stats, the two scenarios are no different.

Europeans share more with Bedouin than Dai do, and share more with Dai than Bedouin do. This is the same scen as EHG btw ANE and WHG. Why this is is equally unclear without formal tools.

@ David

Actually the question is, why shouldn't we try?

Esp when the only thing needed is another Laz model, or ADMIXTURE run, of the same kind that has been done before. And the f4 stat diagram is striking, is it not?

Why shouldn't we indeed?

Davidski said...

WHG, EHG and ANE are part of a continuum across time and space on the mammoth steppe. I don't see the point in making them out to be as different as the largely Basal Eurasian Boeduins and ENA Dai. Like I say, the comparison doesn't work for me.

ryukendo kendow said...

@ David

Let's try to make sure we havve 'cline' nailed down as a concrete concept.

If we think abt what clines do, they make trees fuzzy and merge branches with 'webbing', but do not change the shape.

If EHG is WHG + ANE, Karelian must contribute to Amerindian to the exclusion of Mal'ta. If EHG is a cline with WHG, the model fails. If EHG is a cline with ANE, then WHG is admixed.

If EHG is a cline with both ANE and WHG, that is the same as saying that EHG is ANE+WHG in the language of stats as the gene flow went both ways, and the Karelian --> Amerindian thing, a very unclinelike fudge, has to be used anyway.

The WHG admixed scenario is anyway very cline-like, just that the cline is in Europe, not Siberia.

There is no way of separating EHG in yamnaya from EHG from EHG survvival. But that East Euros havve higher EHG than would be expected from shared drift with CW tells us that the high EHG in NE Europe does not all come from Yamnaya, or for that matter CW.

If we take into account Yamnaya's West Asian-like ancestry, which evveryone has not been paying enough attention to, I expect the picture wwould be quite different for Yamnaya ancestry in Europe. f3 stats with the IE aDNA gives us hints of this already.

That CW covered the whole of E Euro is irrelevant, because the 'resurgence in local ancestry' took place anyway, and not apparently due to increased mixing from the West, because E. Euro are best modeled as Yamnaya +WHG/EHG, while West Euros and others are not.

tew said...

A significant part of the excess ANE in the Middle Eastern lowlands could just be a result of the successive Persian-based empires (Achaemenid, Seleucid, Parthian, Sassanid), when massive movements and settlement of troops (in large part recruited from areas where ANE peaks like the Persian east and south) occurred toward the ME and Anatolia. It doesn't need to be "very" old.

Chad Rohlfsen said...

I think that the model on S8.11 (b) makes the most sense. WHG shouldn't be in EHG, if it is closer to Native Americans (going in at 58% vs 42% ANE). MA-1 could very well be a sister clade to EHG, but with different outgroups. One is ancestral to Europeans and Native Americans, while the other is more ancestral to South and West Asians.

Chad Rohlfsen said...

y haps agree with that scenario more. R and Q, then an I dominant West Euro. Europe is probably 52% outgroup/Near East UP, and 48% sister clade of EHG.

EHG could be both in NA's to the exclusion of MA-1 and also lack WHG. Which it should as when given the choice, Native Americans avoid WHG.

Shaikorth said...

Chad, EHG's fst-distances to modern Turks, North Italians and Sindhis are almost equal, Lezgins are closer than all three, Armenians are noticeably more distant.

One might wonder, if West Asians and Europeans got their "ANE" from separate clades, would it have an effect such as bringing North Italians closer to EHG compared to Turks?

Chad Rohlfsen said...

Shaikorth,
EHG should also be in ENF, if it is in Loschbour. Although the amount of EHG in the non-Basal part would probably be less than the rate in Loschbour.

Chad Rohlfsen said...

If WHG, ENF, and MA-1 have a good chunk of ancestry from the same group, I'd imagine that would help to explain it. That is just another reason to say that EHG does not contain WHG, but is more related to the source.

Marnie said...

Regarding Oetzi and his WHG+EEF signature, I would point out that Stuttgart had a very similar signature. (See the Seguin-Orlando paper, for example.)

From Lazarides 2013: "Stuttgart” (19-fold coverage), a ~7,000
year old skeleton found in Germany in the context of artifacts from the first widespread Neolithic
farming culture of central Europe, the Linearbandkeramik;"

Stuttgart:

1. is in Southern Germany, within a days walk from the Alps.

2. "pulls east" on a PCA plot.

3. mtDNA is T.

4. has about half "EEF" ancestry, like Oetzi.

5. is 1,500 years earlier than Oetzi.

So I will note that this "pulling" east thing was [at least partly] well established in the Early Neolithic, and seems to be associated with regions adjacent to the Alps (admittedly based only two samples.)

Chad Rohlfsen said...

Stuttgart is full "EEF", not half.

Chad Rohlfsen said...

EEF was originally based on her.

Marnie said...

@Chad

"Stuttgart is full "EEF", not half. "

Not in the Seguin-Orlando paper. It depends in which populations and methods you choose, to ascertain components like "EEF", "Middle East", "ENF", etc.

I think a statement like "Stuttgart" is full "EEF"" just goes to show how weak your skills of objectivity are.

You make these kinds of blanket statements all the time. It makes you sound like an idiot.

Stuttgart may be "FULL" or "EEF" in one of your analyses, but not in the analysis of other people.

In any case, on a PCA, Stuttgart pulls even more "eastward" than Oetzi, which is my central point. Stuttgart was an early Neolithic farmer living within a days walk from the Alps. Should raise a few questions about why you haven't spent more thought looking at the central plateau and sub-alpine regions of the Balkans.

If you look at recent papers, its these central plateau and sub-alpine regions in the Southern Balkans where domestication first seems to have occurred.

Marnie said...

So, regarding the above triangle, relative to Loschbour, the shift in Europeans is dominated by an "eastward" shift effected by the Balkan-Asia Minor Neolithic, and secondarily, by a "northward" shift from a Karelian like population.

Doesn't really look like a "mass migration" during the Bronze age [to me].

More like a successful pulling together of "WGH" and "EEF" like populations, which were both mostly Mesolithic Asia-Minor-Europeans.

Regarding "ANE", its a secondary effect, probably originating in some place like Kazakhstan during the UP or Mesolithic. It could have reached Europe by any path and as early as the Mesolithic. However, given its preponderance in places like Finland and Germany, its effect seems mostly to have been felt in Northern Europe.

Chad Rohlfsen said...

EEF was created in the 3population paper. It's you, that doesn't understand. Go back and read it.

Chad Rohlfsen said...

You can't use modern clusters like West Asian, Baltic, Central Asian, or what have you, to describe an ancient individual. It's anachronistic. Come back when you figure it out.

Marnie said...

@Chad

I'm not using "modern clusters". I'm looking at your "triangle plot" above, and referring to the Oetzi and Stuttgart samples in this plot.

Separately, I'm referring to the Seguin-Orlando paper.

Regarding your statement "You can't use modern clusters like West Asian, Baltic, Central Asian, or what have you, to describe an ancient individual. It's anachronistic. Come back when you figure it out" . . .

. . . my husband, MIT PhD, is making jokes about liberal arts types from Harvard [while doing the dishes], noting that "they don't have the ability to self question, perhaps in some, reaching the level of a mental illness."

"They're whole approach is ASSERT AND DEFEND" rather than investigate and search for something approaching the truth.

"They're mired down as application layer propagandists."

"They're not scientists."

Marnie said...

Massachusetts Institute of Technology Mission Statement:

http://web.mit.edu/facts/mission.html

The mission of MIT is to advance knowledge and educate students in science, technology, and other areas of scholarship that will best serve the nation and the world in the 21st century.

The Institute is committed to generating, disseminating, and preserving knowledge, and to working with others to bring this knowledge to bear on the world's great challenges. MIT is dedicated to providing its students with an education that combines rigorous academic study and the excitement of discovery with the support and intellectual stimulation of a diverse campus community. We seek to develop in each member of the MIT community the ability and passion to work wisely, creatively, and effectively for the betterment of humankind.

Marnie said...

Harvard Mission Statement

http://www.harvard.edu/faqs/mission-statement

"Harvard University does not have a formal mission statement."

Chad Rohlfsen said...

For populations from the Yenisei and Ob basins, we find significant evidence for admixture with a variety of Siberian and European source populations (Table S14). The best pair of source populations (i.e., the most negative f3 statistic) involves Swedish MHGs (Motala,Ajv58) and Evens (a northeast Siberian population) (e.g. f3(Shors; Evens, Motala) = -0.012, Z = -9.1). Altogether, these results suggest that contemporary Siberian populations from the Yenisei basin derive part of their gene pool from a Eurasian HG population that shares ancestry with K14, but is more closely related to Scandinavian MHGs than to either MA1 or western European MHGs, indicating gene flow between their ancestors and Scandinavian Europe after K14 but prior to the Mesolithic (36.3 > x > 7 ka BP).

From Seguin-Orlando, in the Kostenki paper. Another confirmation of EHG importance, maybe?

capra internetensis said...

@Chad

Don't feed the troll, man.

It's obvious to everyone who is and who isn't interested in sincere discussion here.

@ryukendo

Thanks for this analysis, it's fascinating. But what could this represent in archaeology? Who was coming from the east and mixing with the predecessors of WHG, and when? Most of W and C Europe post-glacially seems to be Magdalenian, and it's hard to see how they could all have ended up admixed with Mesolithic people from far to the east.

Seinundzeit said...

David,

It seems that the EIGENSOFT suite of applications now includes Fst calculation.

RK,

These comments from the paper are quite pertinent, as far as David's K8 ANE cluster is concerned.

“Next, we added the Karitiana Native Americans, to account for the different relationship between Native Americans to west Eurasian groups, as they share more alleles with MA1 and EHG than with Loschbour. 8). They can, however, be modeled as a 2-way mixture, with successful (|Z|<3) models shown in Fig. S8.9. Thus, for every model we analyzed, our results support ref. 9 that Native Americans are anciently admixed.

In Fig. S8.9a (when MA1 is admixed), Native Americans shown to be a mixture of a population J related to the Onge. Note that this model accounts for the symmetry in the relationship of Native Americans to MA1 and EHG in a different way than that of Fig. S8.6. In Fig. S8.9a, Native American ancestry is more closely related to MA1 than to EHG, however, it is MA1 that derives part of its ancestry from an early node C which “dilutes” its affinity to Native Americans... Fig. S8.9b and Fig. S8.9c present two different solutions (when Loschbour is admixed). Both (b) and (c) agree with (a).”

"The models of Fig. S8.6 have an advantage in postulating that EHG are a mixture of populations related to MA1 and Loschbour, which are actual individuals, while those of Figs S8.10, 11, 12 propose that MA1 or Loschbour are admixed populations, although at present there are no actual individuals that represent some of the admixing populations. Nonetheless, all these models agree on several points: the existence of Basal Eurasian ancestry in Early Neolithic Europeans, the fact that MA1 shares more alleles with Karelia_HG than with Loschbour, but MA1 and Karelia_HG are symmetrically related to Native Americans, and, finally, the fact that the three group of Eurasian hunter-gatherers (EHG, WHG, and ANE) cannot be related to each other by a simple tree, and at least one of them must be admixed.”

I think this means that the non-ENA portion of Karitiana ancestry is from an unadmixed source. So, David's “ANE” cluster is unadmixed (at this scale of resolution), as it is based on Karitiana samples. Yet, despite not being possibly admixed like MA1, David's ANE cluster behaves just like MA1 (in terms of what percentages it yields for contemporary populations, around 40%-43% in Karitiana, 33%-37% in Burusho, 26%-29% in Lezgians, etc). This tells us a lot about the relationship between WHG, ANE, and EHG. If we don't use the possibly admixed MA1, and instead use the non-ENA portion of Karitiana genomes, we get estimates that are still identical to those obtained using MA1. Surely this fact must have some bearing on the possible phylogenetic position of these EHG hunter gatherers.

Although, I do agree that interpreting unsupervised ADMIXTURE output can be akin to tasseography

Marnie said...

@capra

"But what could this represent in archaeology? Who was coming from the east and mixing with the predecessors of WHG, and when? "

I think it's pretty obvious that the Aurignacian, Gravettian, Solutrean and Epi-Gravettian (Eastern Gravettian) population never fully fused.

"EEF", "ENF" or "farmer", Oetzi, Stuttgard, however you want to look at it, looked different from Loschbour.

Moreover, Motala12 already had "ANE" components in the Mesolithic (According to Davidski).

So both the "eastward shifter "ENF" effect and the "ANE" northward shifting effect, were underway or persisted in the Mesolithic.

"Most of W and C Europe post-glacially seems to be Magdalenian and it's hard to see how they could all have ended up admixed with Mesolithic people from far to the east."

I just found a paper on the late Gravettian in Bosnia last night. The site dates to 13,000BP:

http://www.jstor.org/discover/10.2307/529559?sid=21105413818361&uid=4&uid=3739256&uid=2129&uid=3739560&uid=2&uid=70

So it's clearly not the case that "Most of W and C Europe post-glacially seems to be Magdalenian".

And evidence of an Epi-Gravettian in the Upper Volga in the Mesolithic, according to this paper:

http://www.quartaer.eu/pdfs/2010/2010_hartz.pdf

So, capra, I'm truly mystified as to why you would say that only the Magdelanian persisted in the post glacial across Europe.

"and it's hard to see how they could all have ended up admixed with Mesolithic people from far to the east."

The Far East? You mean the Balkans and the Upper Volga?

Marnie said...

Another paper on the Balkans (UP and Mesolithic):

New Palaeolithic discoveries in Albania

Thomas Hauck, Ilir Gjipali, Jurgen Richter, Rudenc Ruka

ABSTRACT The territory of Albania is centred directly within the “Eastern Trajectory” of modern human migrations before, during and after the late glacial maximum. For this reason the Cologne CRC806 project started archaeological fieldwork in 2012 in collaboration with the Albanian National Institute of Archaeology. The most outstanding discovery are several caves in northern Albanian Mati district with undisturbed deposits in Jurassic limestone ridges about 300m a.s.l. Artefacts discovered on the hillslope indicate the presence of Middle and Upper Palaeolithic terrace deposits in front of the caves buried under several meters of Holocene
layers. In a small test trench within the entrance area of one of the sites, called Blaz Cave, we discovered Epi-Gravettian remains. Similar artifacts were excavated 170km to the south in a collapsed cave near Orikum. Together with high-density Mesolithic levels discovered in the site of Shën Mitri (Butrint basin), the CRC806 investigations testified the high potential for future Palaeolithic research in these areas.

Link:

http://www.researchgate.net/publication/261284682_New_Palaeolithic_discoveries_in_Albania

capra internetensis said...

@Marnie

Thanks for the links - but Bosnia and the Upper Volga are not generally considered to be part of Western and Central Europe!

I was specifying that area because our defining WHG samples, Loschbour and La Braña, are from Luxembourg and northern Spain.

KO1, from Hungary, also clusters with Loschbour and La Braña, which is not what you'd expect if there was some kind of EHG-like cline toward the Balkans. We have far too few samples to really say, however.

I am asking questions here, not making any kind of argument, because I know very little about the Late Upper Paleolithic and Mesolithic of Europe. But the influence must have reached all the way across Europe to Spain, if it happened in this period.

Or it could be an earlier thing, say Gravettian, with the other WHG population being western Aurignacian or whatever, in which case what was ANE?

Marnie said...

I just looked and there are apparently quite a number of sites being excavated as we speak in the Southern Balkans showing probable post glacial Gravettian associations.

There are also several sites in Armenia, the subject of this thread, showing Late UP/Mesolithic sites associated with the Epi-Gravettian (Kalavan 1, for instance.)


Mike Thomas said...

Capra your rigt about the HG from the Carpathian
As I stated above ; that's the result of a colonisation of Palaeolithic europe of a subset of balkan foragers via the Hungary.
But who knows what balkan foragers ("BHG") might look like when actually sampled ?

Marnie said...

@capra

"Or it could be an earlier thing, say Gravettian, with the other WHG population being western Aurignacian or whatever, in which case what was ANE?"

It doesn't have to be an "earlier thing."

A late Gravettian and or Epi-Gravettian persisted in the Balkans and Armenian (and maybe in the Upper Volga.)

Many sites in these areas being excavated as we speak.

So no reason to particularly assume that the shifts in Western European populations came from "far to the east".

For instance, assuming that the shifts in European populations in the Neolithic arrived directly from Lake Baikal is unrealistic, especially given abundant evidence for a late UP Gravettian/Epi-Gravettian in the Balkans, Upper Volga and Armenia.

Marnie said...

@Mike

"that's the result of a colonisation of Palaeolithic europe of a subset of balkan foragers via the Hungary. "

I just glanced at a paper this morning suggesting that hunters where following game between Hungary and the Vardar (Macedonia/Greece) in the Late UP. I don't know about colonization, but it doesn't look like UP hunters in the Vardar were cut off from the Danube.

Here's the link:
http://www.paleoanthro.org/static/journal/content/PA20120131.pdf

Marnie said...

@capra

"KO1, from Hungary, also clusters with Loschbour and La Braña, which is not what you'd expect if there was some kind of EHG-like cline toward the Balkans. We have far too few samples to really say, however."

For the Balkans, Alps and Danube, what I suspect is that, as KO1 indicates, riverine areas such as the Danube and Vardar will weight toward WHG.

Ancient DNA of the sub-alpine regions of the Alps, Dinaric Alps, and perhaps Mountain Albanians, where shepherding is viable, will weight toward "EEF" or "ENF".

Mike Thomas said...

Maybe maybe not
But my point was that WHG , especially that HG from Hungary is likely to be a subset of Balkan HGs. Im not confident that the Balkans will be a mere end point of a WHG continuum, but rather be more diverse ; and perhaps form its own "BHG" as dos sweden .

Davidski said...

Swedish foragers are 15-19% ANE. That's the only reason they make their own SHG cluster.

KO1 from Hungary is 0% ANE. So I'd say it'd be a miracle if Balkan foragers had any ANE.

Mike Thomas said...

But what if the Balkans also had EHG from the ukrainian refuge ?
I don't think you can use Hungary as a direct proxy for the Balkans- that's been my point

Colin Welling said...

@David

Most of the non-Basal Eurasian input in modern Near Easterners is very WHG-like, so much so that it's difficult to differentiate from WHG. So why would foragers in southern Italy and the southern Balkans carry a component that was distinct from what was all around them in Europe and the Near East? It wouldn't make any sense.

Could you explain to me again the basics on the distribution and time scale for the WHG-like component? I'll explain what I know and please correct me if I'm wrong.

WHG and EHG were already differentiated by the Mesolithic, as demonstrated by Samara HG vs Brana. WHG clearly derives LGM Europe given its near absence in modern Middle Eastern people and the dramatic drop in percentages of WHG in farmers from Southern Europe.

WHG may be admixed (West European LGM + East European LGM?) or EHG may be admixed (European LGM + Central Asian/Siberian LGM?). Whatever the case, early European farmers carried the WHG signature which they surely picked up somewhere in Europe; the authors of the first Laz paper suggest the Balkans. But you claim there is another WHG like component which is different from lo./brana and is from the Near East. Why?!? That would potentially mean there is Brana-WHG (found), balkan-like-WHG (suggested by authors to explain mixed farmers), some mysterious West Asian-like-WHG (only suggested by you?). How does all this jibe with the authors in the first paper saying that WHG is absent in West Asia? Where and when was this "West Asian-like-WHG" prominent in West Asia and how did it get into early European farmers. What do you think was the composition of the earliest farmers in greece?

I find all this extremely relevant to determining the what Southeast Europe was like during the mesolithic. If WHG-like stuff was in the mesolithic middle east then it was surely strong in the mesolithic balkans and that makes it a little less likely that the southeast european mesolithics were significantly different from the western european mesolithics. IF, WHG-like was not in the middle east then we don't have the same restriction as above and given how West europe was different from east europe during the mesolithic its certainly possible that southeast europe was different too.

Colin Welling said...

But what if the Balkans also had EHG from the ukrainian refuge ?
I don't think you can use Hungary as a direct proxy for the Balkans- that's been my point


European farmers, who came from the greece and the balkans, by and large lack ANE. The balkans did not harbor ANE. Im of course ignoring the fringe which is irrelevant here.

Matt said...

Chad: From Seguin-Orlando, in the Kostenki paper. Another confirmation of EHG importance, maybe?

Yeah, that's well remembered. I think it's another indication that EHG seems to give a stronger signal contributing to populations near Eastern Europe (Yenisei and Ob basins) compared to MA1 and a comparable signal to MA1 to Amerindians.

Sein:If we don't use the possibly admixed MA1, and instead use the non-ENA portion of Karitiana genomes, we get estimates that are still identical to those obtained using MA1. Surely this fact must have some bearing on the possible phylogenetic position of these EHG hunter gatherers.

To me, I find it all pretty difficult to interpret all this.

I mean in the models in the paper where MA1 is admixed, the "ANE" node contributes 24% to Karitiana and 31% to MA1 (while when Loschbour is admixed the "ANE" node goes 60% to Karitiana). As opposed to the 40% of the ANE cluster in Karitiana in David's ADMIXTURE.

The whole idea of ANE is only based on formal statistics finding MA1 as unadmixed. If these same formal models later find MA1 is admixed based on a new sample, what is the grounding for the idea of ANE?

That ADMIXTURE Eurogenes K8 run and cluster was and is interesting because it supports the Laziridis 2014 formal model, rather than that it necessarily gives the best fit to the populations involved. So if the Laziridis 2014 formal model for whatever reason doesn't work, it's difficult to know what to make of it.

Colin Welling said...
This comment has been removed by the author.
Mike Thomas said...

My main point was that the Balkans might have had a different structure to WHG . Whether it had ane / EHG is not the main issue for me

Colin Welling said...

@chad

I think that the model on S8.11 (b) makes the most sense. WHG shouldn't be in EHG, if it is closer to Native Americans (going in at 58% vs 42% ANE). MA-1 could very well be a sister clade to EHG, but with different outgroups. One is ancestral to Europeans and Native Americans, while the other is more ancestral to South and West Asians.

Hu, interesting concept. Maybe that would explain the early forms of r1 stuff in both the middle east and europe. then the later r1a/r1b stuff that came from europe was like a second ane expansion into the middle east carrying newer r1a/r1b.

It would be interesting to thing that europeans and middle east share two layers of ane.

Davidski said...

The K8 ANE is a stable genetic signature in modern populations, like the West African and East Asian clusters that we always see in ADMIXTURE runs, except it's more difficult to isolate, probably because of recent drift.

It's certainly relevant to the study of major population movements, even at a very fine scale, like in the post above.

So this is all too much of a coincidence if ANE is just some mirage. But I'll probably have more to say on the matter after I've seen the EHG and Yamnaya samples.

Colin Welling said...

My main point was that the Balkans might have had a different structure to WHG . Whether it had ane / EHG is not the main issue for me

I agree with your point, then.

I actually lean towards southeast europe being different from west europe during the mesollithic., whether that is because of Basel Eurasian or something different that just hasn't been described.

Colin Welling said...

much of this depends of whether basel is real and if the estimates of basel are correct.

Davidski said...

Colin,

All I can say is that Near Easterners are partly WHG-like, and this isn't because of admixture from Europe (although there is extra WHG in parts of the Near East which might be due to European admixture).

This can't be mostly EHG, because if it were, then the least ANE and Sub-Saharan admixed Near Easterners wouldn't be pulling towards the WHG/ENF axis; they'd be pulling east. It just looks like to me that they're lacking ANE/EHG type stuff, and their Neolithic ancestors probably had 0% of it, just like European Neolithic farmers.

My bet is that what they have is proto-WHG stuff that entered Europe from the Middle East before the Ice Age.

Some of the recent comments here about how highly differentiated the West Eurasian and Siberian foragers are seem really strange to me. Considering the time and space, they appear to be closely related. Also, clearly there was continuity in the north because we have MA-1 and AG-2 in Siberia from either side of the Ice Age peak both belonging to the ANE clade.

But like I say, I need to see how the EHG and Yamnaya behave to say much more.

Marnie said...

@Mike Thomas
@Colin Welling

"My main point was that the Balkans might have had a different structure to WHG . Whether it had ane / EHG is not the main issue for me"

I also agree with this point.

I'm also trying to suggest, based on my own observations of the geography, people and history of Albania, Epiros, Macedonia, Upper Greek Western Macedonia, Lower Greek Eastern Macedonia, Greece and Bulgaria, that there may still be genetic variability in different pockets of the Southern Balkans and Greece.

So, as Mike has stated, we can't extrapolate what the Balkans might have been like in the Neolithic or Mesolithic from one sample. Frankly, I think the Balkans is really complicated and will take about as much coverage as there is for Germany in the Haak paper. (maybe more)

Here's a map showing the mountain ranges of the Southern Balkans:

http://www.apan.gr/contents/uploads/image/Fragments%20p_83%20%20(gia%20site).jpg

Until the Egnatia Odos (Via Egnatia Motorway) was constructed between 1994 and 2009, it used to take more than ten hours to drive from Salonika to Corfu. Fifty years ago, it was barely more than a horse path, so you can get an idea why there could still be some genetic stratification in the Southern Balkans.

There's also the different geography to consider, with some areas being Mediterranean in climate, some areas having a continental climate, and even some remaining glaciers in the Dinaric Alps.

Anyway, I'm just trying to illustrate why one would need to be careful when trying to extrapolate what the Neolithic in the Balkans might have looked like from a genetic perspective.

Colin Welling said...

ya, what looks pretty clear is that most of the migration in north eurasia from the time of MA to the neolithic seems to be along the east west axis. hence the continuum of WHG-EHG-ANE.

I actually think this is makes a stronger case for EHG being from north eurasia, rather than the AG guy.

Mike Thomas said...

Marnie
Where did u get that oetzi is East shifted?
To me he looks like he's in the centre of the neolithic clump ?

Mike Thomas said...

Dave
" Also, clearly there was continuity in the north because we have MA-1 and AG-2 in Siberia from either side of the Ice Age peak both belonging to the ANE clade."

I think this is only an apprent continuity in light of what I and Ryu have argued

Colin Welling said...

"So, as Mike has stated, we can't extrapolate what the Balkans might have been like in the Neolithic or Mesolithic from one sample. Frankly, I think the Balkans is really complicated and will take about as much coverage as there is for Germany in the Haak paper. (maybe more)"

Im not saying there was a lot of diversity in the balkans during the mesolithic. If there was it was between an WHG type in the north and an EEF type in the south. I don't think there can be anything else.

Though, i must admit. I wonder what the hell mesolithic italians will come out as. they may have been like an island even in the mesolithic.

Marnie said...

@Davidski

"Some of the recent comments here about how highly differentiated the West Eurasian and Siberian foragers are seem really strange to me. Considering the time and space, they appear to be closely related. Also, clearly there was continuity in the north because we have MA-1 and AG-2 in Siberia from either side of the Ice Age peak both belonging to the ANE clade."

From the archaeological record, as well yDNA hgs, and autosomal DNA, there definitely is a close relationship, as well as some sort of continuity, but in the details, there's a lot of history, almost 20,000 years worth.

Davidski said...

Mike,

MA-1 and AG-2 form an autosomal clade and both belong to mtDNA U. This looks like "apparent" rather than real continuity to you?

More likely, you and rk are hoping and fishing for something that isn't there.

Marnie said...

@Mike Thomas

"Where did u get that oetzi is East shifted? "

compared to Loschbour and Sardinians, is what I mean.

If Oetzi was already somewhat admixed, then his ancestors, the non-WHG ones, would have to have been shifted even further to the "east" than Oetzi.

Marnie said...

@Colin Welling

"Im not saying there was a lot of diversity in the balkans during the mesolithic. If there was it was between an WHG type in the north and an EEF type in the south. I don't think there can be anything else."

More or less. I just read this morning in a recent paper on faunal analysis that they suspect that Upper Paleolithic hunters were following rivers between the Danube into the Vardar. So the Vardar would maybe in the UP look more WHG like.

I agree though that Southern Greece by the time of the early Neolithic must have looked more EEF like.

Which is why I say it's complicated.

ryukendo kendow said...

@ Mike
There was continuity from mal'ta to AG-2. More generally, the distr of autosomal branches began to look like its present distr around the LGM.

@ Davidski

That continuity is from 24 kya to 17kya, and is bracketed by discont from Ust-Ishim and Kostenki on one side and NAms and ENA Siberians on the other.

So the idea that K got to Siberia, then incubated there for 25kya to give rise to all else incl. R, which is an idea that some may get from your comment above, isn't very likely.

Look forwward to you getting the genomes.

Marnie said...

@Colin Welling

"Though, I must admit. I wonder what the hell mesolithic italians will come out as. they may have been like an island even in the mesolithic."

I don't know. It will be fascinating to find out.

I do think there must have been some continuity between Italy, along the Adriatic Coast, down to Albania and as far as Corfu, in the Upper Paleolithic and Mesolithic.

I'm basing this guess on sea level, as well as some preliminary sites in Albania that look Gravettian like.

Krefter said...

@Davidski,
"MA-1 and AG-2 form an autosomal clade and both belong to mtDNA U. This looks like "apparent" rather than real continuity to you?"

Did you conclude AG2 has U on your own analysis?

ryukendo kendow said...

@ All
As an aside, for the point that EHG survvival might influence the Yamnaya estimates, I posted two pics. Please pay more attention to the second one, not the first.

http://imgur.com/kpNrdcX,k73e5qN#1
Once again, NW Euros orange, NE Euros green, others black. Yamnaya shifts west, EHG shifts south.

Chad Rohlfsen said...

Matt,
I certainly find EHG more important than ANE, whatever it might be. The fact that EHG and SHG are more related to Native Americans than MA-1, is astounding. It definitely makes sense considering the back and forth across Eurasia. I think it was the LGM that caused a mass homogenizing of WHG, making it look more unadmixed than it really is.

ANE could be a real signal, but what exactly that is, is the question. Maybe MA-1 and EHG are two branches, but ANE/MA-1 is more related to Central, South, and post-Neolithic West Asians.

Maybe, Yamnaya is 50%EHG, 35% Near Eastern and 15% MA1 (which could be dominated by EHG)

Modern N Europeans might be in the 40% EHG range.

I'm really crossing my fingers to see if David can pluck apart WHG and MA-1 with a quality EHG genome. There is some diversity within WHG, as the scatter plot and k-stats had Loschbour as closer to EHG.

Davidski said...

Corded Ware had more EHG than Yamnaya and any modern population.

This might be because Corded Ware and Yamnaya split from different Proto-Indo-European groups that existed before Corded Ware and Yamnaya, and had different levels of EHG.

I'm thinking the group with the higher EHG might have been Sredny Stog, which was identified as part of the Proto-Indo-European horizon by Mallory. On the other hand, the Samara Yamnaya nomads might have come from Khvalynsk and/or Repin people, who maybe had higher levels of Near Eastern and perhaps even some Central Asian ancestry.

Mike Thomas said...

Corded and yamnaya, whilst certainly entertaining LN - EBA contacts, share a common EHG ancestry which predates and has nothing directly to do with Mallory's bogus "proto-IE" Repins and Sredny Stog.

I suspect that only when you (collectively) realize that the genetic model constructed by reich et al fails you'll you'll bother to endeavour a 21st century understanding on the fact that the pre-Yamnaya groups of the steppe weren't "proto"-anything; but separate and diverse steppe forager groups .

ryukendo kendow said...

CW might have more EHG than Yamnaya, but that still doesn't explain anything as CW do not share the highest drift with present-day euros with the highest EHG. Rather, CW and all its descendant cultures as well as Yamnaya itself show a pattern of having most drift with NW and NC euro pops, despite differing in the same way from all present-day Euros(high EHG) as how E Euros differ from all other Euros(high EHG), making me think this is because E.Euros do not havve the West-asian portion of both CW and Yamnaya that strongly, while NW Euros have both the EHG and West-Asian portion strongly, suggesting less CW and Yamnaya ancestry for E Euros than is usu assumed.

@ Davvidski
Where did you get that CW has more EHG than Yamnaya? I don't see a stat in the paper to that effect.

Davidski said...

My point was that EHG was associated with the westward Kurgan expansion at a very early stage, and not a post-Kurgan expansion within Eastern Europe of EHG like rk is claiming.

By the way, we already have mtDNA from Neolithic Ukraine and Ukrainian Bronze Age Kurgans, and a lot of it is mtDNA C, like what the R1a Karelian forager had.

http://eurogenes.blogspot.com.au/2011/09/surprising-adna-results-from-neolithic.html

postneo said...

@Davidski
AG-2 and MA-1 show local continuity but their descendants did not move west till well after the ice age.

ryukendo kendow said...

That wasn't my point, was it now?

Davidski said...

rk,

Actually, what the paper says is that the ancestors of the Corded Ware sampled in Germany probably had higher EHG than Yamnaya. Page 116.

"This hints that the Yamnaya population, while visibly a much better ancestral source for the Corded Ware, may in fact not be the exact admixing population. This also agrees with the results of N=3 discussed below and presented in Extended Data Fig. 3, which estimates that the Corded Ware can be modeled as 29.1% Esperstedt, 9.4% Samara_HG, and 61.5% Yamnaya, which suggests that the population of eastern migrants had a slightly higher proportion of EHG ancestry in its makeup than the Yamnaya sample from Samara."

The useful thing to keep in mind here is that the Yamnaya horizon was divided into six distinct zones based on archeology, which might mean there were genetic substructures within it too.

Another point to keep in mind is that the Samara Yamnaya nomads were probably not the direct ancestors of Europeans, but rather closely related to other Yamnaya and pre-Yamnaya Indo-Europeans who were.

Mike Thomas said...

I can't wait for aDNA from polish globular amphora and other late neolithic & residual Mesolithic groups .

Davidski said...

GAC genomes won't be very interesting. They'll be clones of Oetzi the Iceman and CO1.

I can't wait for aDNA from Polish Corded Ware and those Ukrainian Kurgans with mtDNA C samples.

My bet is we'll see higher EHG than in the German Corded Ware and Samara Yamnaya.

Grey said...

@capra

"Or it could be an earlier thing, say Gravettian, with the other WHG population being western Aurignacian or whatever"

It always seemed unlikely to me that the farmers would derive from their own isolated HG population so WHG and EEF more likely represented two separate HG populations one or both of which developed farming later and expanded dramatically as a result and this earlier structure might partially resemble the later farming expansions.

For example if HG A were initially widespread Aurignacian in the west and HG B was Gravettian intruding in a wedge through Anatolia and Central Europe then you might end up with a train in a tunnel type distribution of the two types of HG with HG A as the tunnel sides and HG B as the train.

If the Levant was part of the tunnel distribution and a farmer A population developed out of the HG A population there and expanded with a mostly maritime spread it would largely overlay HG A and if at a similar time a farmer B population developed somewhere in the HG B distribution and followed the same Anatolia -> Central Europe route as HG B then farmer B would largely overlay HG B.

(With a lot of criss-crossing and mixing of course but that underlying structure.)

.

"in which case what was ANE?"

mountain giants

Scandinavia, Urals, Caucasus etc

Mike Thomas said...

I'm hypothesising that corded ware more or less "derived " from preexisting Mesolithic and Late Neolithic groups of GAC and Ertrobelle type cultures, as well as late Tripolye. Naturally some admigration from yamnaya occurred, however even more importantly was gene flow from west Asia and Tripolye into yamnaya which transformed an otherwise diverse and rather unremarkable set of forager groups .

Then, the ANE shift in Central Europe was part of what has been termed the " HG resurgence", eastern variant thereof; with secondary colonisation of Western areas like germany by their "polish neighbours" rather than wholescale Yamnaya replacement .

The ANE - R1b M269 shift in Western Europe, will also beg another explanation . I think it's more complicated that Yanaya hiking up the danube. A not insignificant part of ANE in western eruope might come about from a northern direction rather than directly from the Carpathian - Hungary. Ie the Low Countries - given their role in initiating the BB phenomenon


But the absence of Bronze Age specimens from western Europe precludes conclusions as to how sudden the ANE shift was

Davidski said...

Your rather complicated attempt to explain away the Kurgan expansion ignores the obvious fact that most European males today carry Y-haplogroups that reach high frequencies in Kurgan remains, R-M417 and R-M269, which expanded very rapidly during the Copper Age thanks to intense founder effects, also sometimes known as the Genghis Khan effect.

Krefter said...

@Mike,
"The ANE - R1b M269 shift in Western Europe, will also beg another explanation"

IMO, the bronze age Celtic-spread proposed by Maciamo and many members at Anthrogencia makes a lot of sense.

"Celtic" origins are based on outdated 1800s research and has been restricted to the Hallstatt, La Tene, and Classical records. The spread of Celts is certainly much older.

P312 in bell beaker all the way back some 4,3000YBP, R1b in Urnfield 3,000YBP, and a very derived British-Irish clade of L21 in two British Celts 2,000YBP is evidence of this IMO.

Modern L23 obviously show L11 arrived from the east probably after the middle Neolithic, and stormed W. Euro very quickly.

The quick spread is supported by the distinct geographic distribution of L11 and P312 clades.

L21 is North Sea
Df27 is Atlantic
U152 is Central-Italian
U106-Df100-Df99-Df19-L238 are central-northern.

This supports a bronze age or earlier spread IMO.

Think about it. L21-rich north sea people were Celtic, U152 rich Central-Italians were Celtic and Italic, Df27 rich Iberians were mostly Celtic.

Add to this the clear Germanic motif of U106 and Z2105 and L23* uniformity in Samara Yamna.

No one can use the argument "Well Basque and many pre-Roman Iberians and French were/are non-IE" anymore because Haak 2015 proves there's alot of Yamna/EHG-type ancestry in all SW Euros.

If Celts mostly spread lets say 4,000-3,000YBP, Celts in Roman times would be like Slavs today. I can't imagine proto-Celtic being from Hallstatt. That makes them almost as close as Americans are to Shakespeare.

Chad Rohlfsen said...

Mike,
Those Beakers are at the beginning of the Bronze Age. These guys go to the NW and SW. The oldest bronze dagger in Britain is 2200BCE, and already fully British in style.

Krefter said...

I watched a very recent documentary on the history of the Celts and read a recent book on it. Their only reason for saying Celts spread with Hallstatt and La Tene is because those were the Celts Classical writers knew, and therefore the oldest confirmed Celts. This doesn't leave out the possibility that there were Celts in 1,500BC.

The doc used the "lack of central European ancestry in Iberians" as evidence Celts spread by contact. In reality Iberians look to be about 50% R1b-rich German Bell beaker and 50% Spanish Middle Neolithic.

Nirjhar007 said...

@David
Off topic what is your thought on upcoming Australia Vs. New Zealand in the 28th? Both look in Great form!

Krefter said...

@Nirj,

Baseball is the real shizz.

https://www.youtube.com/watch?v=vYMhT_wgBwE

ryukendo kendow said...

@ David


The keyword here is 'slightly'. That is very different from saying CW has more EHG than Yamnaya. It still prob has less.

But anyway, as I have said before, this does not explain anything. Zilch. CW could be higher EHG than yamnaya and still it does not explain anything.

We already know that both CW and East Euros have higher EHG than others, incl. possibly Yamnaya. The key question here, is why does Corded Ware--not yamnaya, Corded Ware--seem to share highest drift with NW Euros, and not E Euros, if this is the only thing that mattered?

Why would all the desc cultures of CW display this tendency to favour NW Europeans over E Euros? Why would a comparison of CW vs Yamnaya and CW vs Karelian show that NW Euros are closest to CW in Yamnaya ancestry, while NE Euros are closest to CW in EHG ancestry, which is what the f4 stats show?

The answer obviously is that Yamnaya, but more importantly even CW, had high levels of a type of ancestry that we cannot place a finger on now--no aDNA, but we can already gauge. This ancestry was from the NE and Cauc which is preserved well among NW and NC euros, but somewhat less so among E Euros, who might actually have less CW and Yam ancestry than NW Euros, but have their high EHG inflate similarity with both Yamnaya and CW to some degree.

Note that this pattern is mediated strongly through CW, not through Yamnaya, so CW is also party to this, whatever the ancestors of CW were.

This is also borne in the ADMIXTURE analysis. At K=20, both Yamnaya and Corded Ware are strongly West Asian, as well as all the IE cultures, more so than any European today; and NW and NC euros are too, but NE Euros, esp finnish, lithuanian, estonian, have almost none.

Consider the implications.

«Oldest ‹Older   1 – 200 of 412   Newer› Newest»